Gopher.ICU - Computing

Sinclair-ZX-Spectrum-48k

2021-12-21T19:06:40+01:00

# Sinclair ZX Spectrum 48k+ ## Child of the 80's Like many children of the 80's my first computer was a Sinclair ZX Spectrum. I spent many hours on it playing games, bought magazines, books and even tried my hand at programming. Programming for me at that time meant typing out long pages of machine code that often resulted in several wasted hours. Rarely did it work... ## The motivation It's been a strange week. Somehow the topic of Z80 computers came up and I started doing a little digging. I actually had no idea, or had forgotten, that the humble ZX Spectrum had a Zilog Z80 processor at its heart. Remembering that I still had one complete Sinclair Spectrum 48k+ and a case in it's original box (from my original Spectrum) in the loft, I felt motivated to do something with them. Sadly a day or two later came the news of Sir Clive Sinclair's passing. The man who had inspired a generation of computer enthusiasts with his accessible home computers was gone. The news both saddened me and strengthened my resolve to bring my Spectrum back to life, as a personal tribute to this great man. ## Bringing it back to life My first task was to get the computer and box out of the loft. To my surprise there was also an original Boots cassette player up there too. Unfortunately on inspection the complete Spectrum wasn't in great shape as it had been left in a damp loft for many years. The case that was inside the box however was just as fresh as the day I put it in there in the late 80's or early 90's. The reason the case was empty was that I had bought a better quality case for it with a PC style keyboard. I cleaned off the main board[1] as best I could and took note of the issue number so that I could order a replacement capacitor kit. After doing some reading up and watching youtube videos detailing the restoration process, I ordered a capacitor kit. This kit[2] included a full set of replacement capacitors, a heatsink for the ULA and a 100uf capacitor to perform a composite mod. While waiting for the kit to arrive I tried to assemble the various bits of Spectrum hardware I had laying around. I managed to find a power-play joystick interface and a Currah Speech interface along with its manual. Unfortunately I couldn't find the PSU or cassette cable... *WARNING* - If you have an old computer such as this, do not attempt to power it up without first changing the capacitors. It is likely they have dried out and are way out of specification. Attempting to power up the system could result in destroying other sensitive electronic components on the board or even a fire. ## Capacitor replacement Replacing the capacitors was a little fiddly but I knew what to expect after watching several youtube videos. I did the composite mod first, then I replaced the capacitors one by one, making sure I replaced each with the correct value and in the correct polarity. The whole process took me an hour or so to complete and I was fairly satisfied with the result[3]. ## Cleaning and reassembly I spent some time cleaning up the edge connector with a fibreglass pen and removing any remaining solder flux. The board was then given a light scrub with contact cleaner. On checking my original case, everything was in good order and just required a light clean inside and out before beginning reassembly[4]. The computer is now back in one piece[5] and awaiting a modern switching PSU to arrive so that I can test the system. ## PSU arrived (20/10/2021) It doesn't work... The power supply arrived about a week ago but I was too disappointed to do anything more with it. The past few days I have been investigating the issue with a multimeter and reading up online. It seems that a common fault with these old spectrums is the lower memory modules become bad, either with age or from other issues that affect the voltages going to the chips. The result is garbled colours on the TV or black & white bands and other vertical lines. ## Lower Memory The lower memory consists of 8 x 4116 chips[6] and requires 3 separate voltages (-5v, +5v and +12v). If any one of the voltages goes too far out of specification it can damage the chips. Having checked all the voltages everything seems to be within specification, so I can only assume they have just expired through age. From what I have read they are quite prone to failure and also becoming increasingly hard to come by. Thankfully there are options to replace the problematic 4116 chips: 1) Purchase an all in one Lower RAM replacement board[7] 2) Use 4164 upper memory modules Both options require removal of all 4116 lower memory chips. ### Lower RAM replacement board The easiest, and cheapest, repair is option 1. The board requires soldering only the necessary legs from the daughter board to function, rather than all 16 legs of every replaced chip. It does however not look pretty, IMHO, so if you are hoping to keep your spectrum looking original then you may want to consider another option. It does use modern chips and a single voltage however, so it should be more energy efficient, generate less heat and be more reliable. ### Use 4164 memory modules Using 4164 chips, which seem to be more robust and available, is an option I considered. Unfortunately it works out more expensive and is quite a bit more work. It would however retain the aesthetic of the original unit. If I had gone this route I would have socketed all the lower memory, so that future replacement would be fairly trivial. The 4164 uses only a +5v feed and so requires some modification to the chips, or the board, to remove the unnecessary voltages and supply +5v to pin 8. I wanted to avoid making many changes to the board and make the changes easily reversible. The 4164 requires no -5v to pin 1. Looking online people have solved this by simply cutting off the leg of the chip, this saves breaking any tracks on the board, but obviously you are damaging the chip and preventing later use for some other application. The +5v to pin 9 can remain in place, but we need to provide a +5v feed to the pins which provided +12v to the 4116 chips. The only neat way I could see of doing this was to cut the 12v supply track somewhere near the positive terminal of capacitor 44[8] of my model 4B board. Then a short bridge wire could be installed between the nearby +5v and previous +12v rail to provide the required power for the chips. *NOTE* - The above 4164 modification is speculative. This modification is untested by me and I have no source proving it would work. ### Removal of 4116s (23/10/2021) Knowing this was going to be tricky, I ordered a good quality solder sucker to make the job a little easier. Even with the premium solder sucker it took me the best part of 3 hours to carefully remove the lower memory[9][10] without damaging the chips. ## Ordered Parts (24/10/2021) Placed order with retroleum.co.uk for: 1) lower memory module 2) switching power regulator 3) mono cassette cable 4) 8x 16 pin IC sockets ## Parts arrived (30/10/2021) I was quite excited to receive the package in the post this morning, so I quickly got all the soldering gear out and set to work. I replaced the power regulator first, followed by the memory module, taking care not to get the legs too hot. Unfortunately, upon testing, there is still a problem. The thick black and white bands have now gone, but the coloured blocks that are alternating background and foreground colour remain. I don't know what the problem is, practically everything has been replaced. The only notable thing is that TR7 is rather hot and there is a quiet wheezing noise coming from the speaker... At this point I have put around 40GBP and 6 hours of my time into restoring it and I'm about all out of enthusiasm to invest any more time or money. ## Ordered Parts (09/11/2021) Placed order with retroleum.co.uk for: 1) 1x 28 pin DIL socket 2) 1x ROM replacement module After emailing Phil at retroleum.co.uk and explaining my continued issues he suggested that it would be best to try identifying the problem with a diagnostics ROM, rather than blindly changing components. For less than 10 pounds, including postage, it seemed like a sensible thing to do. The side benefit of this approach is that the ROM module has a switch and can be programmed with two ROM images. I chose the standard Sinclair 48k ROM and his diagnostics ROM ,so if there is a problem with my original ROM the new one can be left in place. ## Parts arrived (12/11/2021) I took my time desoldering the ROM module and then proceeded to solder in the DIL socket. The DIL socket will allow easy replacement of the ROM in future. Once the socket was installed I fitted the ROM card and connected the board up to a TV and PSU. To my surprise the Sinclair copyright logo appeared! It would seem that the issue was with the original Sinclair ROM. I switched it off and connected up the keyboard but unfortunately only some of the keys worked. At this point I knew the system was working and needed a replacement keyboard membrane. I figured having gone this far I should get one ordered. ## Ordered Keyboard Membrane (24/11/2021) After reading on the retroleum.co.uk website why Phil doesn't supply spectrum + membranes and reading on his blog that he had a good experience with one provided by RAWP. I proceeded to order one from them via sellmyretro.com[11]. Unfortunately not the cheapest option at almost 20 pounds, including shipping. ## Membrane arrived (03/12/2021) I fitted the membrane without issue and on testing all keys functioned as they should, result! I installed the 'ZX Tape Player' app to my android phone and, with the volume turned up to full, managed to load and play a game or two of 1943 & Jetpac. [1](gopher://gopher.icu/I/phlog/Computing/images/spectrum/before.jpg) [2](gopher://gopher.icu/I/phlog/Computing/images/spectrum/refurb-kit.jpg) [3](gopher://gopher.icu/I/phlog/Computing/images/spectrum/after.jpg) [4](gopher://gopher.icu/I/phlog/Computing/images/spectrum/base.jpg) [5](gopher://gopher.icu/I/phlog/Computing/images/spectrum/complete.jpg) [6](https://sinclair.wiki.zxnet.co.uk/wiki/DRAMS) [7](http://retroleum.co.uk/zx-spectrum-chips) [8](https://spectrumforeveryone.com/wp-content/uploads/2017/08/ZXSpectrumIssue3b-Layout.gif) [9](gopher://gopher.icu/I/phlog/Computing/images/spectrum/lower-ram-fr.jpg) [10](gopher://gopher.icu/I/phlog/Computing/images/spectrum/lower-ram-bk.jpg) [11](https://sellmyretro.com)

My-next-hardware-project

2023-02-14T21:34:14+01:00

# My next hardware project Over time I have collected a lot of old hardware. It was all in working order at some point, but over time batteries and hard disks have died or they have been cannibalised for other projects. ## The contenders ### Unnamed 4U Dual Pentium Pro 200mhz ATX motherboard, 128mb RAM. It will hold standard ATX components and PSU. Possible RAM upper limit of 256mb... I took 128mb out of it for the SGI. ### Compaq Proliant 6U(ish) Dual Pentium Pro 200mhz (with second daughter board to hold another pair), 384mb? RAM, Compaq Smartraid 2 SCSI card with 4x 9.1Gb SCA SCSI disks. The turn-off for me on this is that a lot of it is non-standard form factor. Picture of both systems[2] stacked. ## Parts and thoughts This week I purchased an 8Gb CF card and a CF card to 40 pin IDE adapter with bracket[1]. I had the idea of resurrecting the 4U dual pentium pro, using it with a version of Linux that will run in memory and maybe a separate CF card with FreeDOS on it. I like the idea of them being switchable ROMs of sorts so may include a separate drive of some kind for user data. Today I was talking to someone on IRC who has 3 Proliants and I started thinking that there was a lot more scope for expansion with the Proliant. I can add a lot more memory and a second pair of CPU's. The SCSI sub system could be removed and replaced with something more modern, maybe a pair of SATA drives... I already have a 3ware hardware raid card and a couple of 200Gb SATA drives in hot swap caddies. Now I'm really in a bit of a dilemma! Thankfully there is a democratic way to settle this since I finished the gopher poll system. Cast your vote![3] [1](https://www.ebay.co.uk/itm/364016036247) [2](gopher://gopher.icu/I/files/which-server.jpg) [3](gopher://gopher.icu/1/poll?poll=1676405525)

Tokenizer

2020-12-11T19:19:59+01:00

# Tokenizer I had the idea after seeing an IRC channel topic asking people to use a pastebin with an example using netcat: `command | nc somepastbin.com 9999` ## Flexibility Initially I thought about just copying the pastebin functionality but then I realised it could easily be used for URL shortening too. The basic premise is to pass some data to the server in return for a token, which can later be exchanged for the stored data. There are a number of options to extend functionality: - Decide how long the data is retained - Count how many times the token is accessed - Limited access tokens (disable/remove the token after X clicks). ## Implementation I wrote a perl module, Token.pm, which handles storing & retrieving of data and token creation. I then created a basic server to listen for data on a specific port which returns a URL containing the token in reply. If the data passed begins with http:// or https:// then it returns a URL which provides a redirect to the URL. Otherwise it returns a URL to view the data as plain text which can either be retrieved with curl or by viewing the URL in a browser: Example: `#echo "This is the cats meow!" | nc MY-IP-ADDRESS 2020` `https://MY-IP-ADDRESS/V/yjOLc2` `#echo "http://google.com" | nc MY-IP-ADDRESS 2020` `https://MY-IP-ADDRESS/R/ztbvx2` The final part was to create a cgi script that would return the data or perform a redirect when viewed in a browser. Ideally this should be hosted on a very short domain name. `#curl -s http://MY-IP-ADDRESS/V/yjOLc2` `This is the cats meow!` ## Status At the moment it works locally, providing URLs for viewing / redirect. It also stores token creation date and counts the number of times it has been accessed. UPDATE 11-12-2020: I have extended the functionality to include viewing with gopher and keeping a count of how many times a token/link is accessed. I also shortened the https:// URL's as much as I could but ideally need a short cheap domain name.

kempston-joystick-interface-repair

2021-12-23T16:47:30+01:00

# Kempston Joystick Interface Repair This was part of an Ebay lot. I was actually only interested in the zipstick joystick that was part of the bundle, but seeing as the bundle was less than a zipstick alone. I figured it was worth a punt, even though the items were listed as 'untested'. The interface is an original Kempston Joystick Interface+ in its original box, complete with the instruction card. Unfortunately on plugging it in to my spectrum any game I enabled the kepmston joystick on would steer left and not recognize any other direction, whether a joystick was plugged in or not. I concluded, after reading some articles online[1], there was an issue with the logic IC. The chip is a DM74LS366AN which no longer seems to be produced. i Luckily I managed to find one NOS on Ebay for 2.50 GBP including postage. I have removed the logic chip and prepared the board for the replacement in anticipation of its arrival. Desoldering the IC with a good quality solder sucker and cleaning up with solder wick was straight forward[2]. ## Update 17/12/2021 Chip arrived and soldered in. Unfortunately no change in behaviour so I've ordered a replacement 74LS138N. It's the only other chip on there so I'm hopeful this one will resolve the issue! ## Update 23/12/2021 Chip arrived and soldered in. Again, no change so I'm stumped. One possibility is that, after testing my new game loader which is fitted with a diagnostics ROM, M1 is reporting as being defective on my CPU. From what I understand this is fairly common on zx spectrums and can lead to external devices not functioning or functioning incorrectly. It looks like I will be replacing the z80 CPU in the near future... [1](https://oldmachinery.blogspot.com/2014/01/joysticksnstuff.html) [2](gopher://gopher.icu/I/images/kempston1.jpg)

Computer-Input-Devices

2021-08-04T08:30:37+02:00

# Computer input devices Over the past five decades computers have decreased in size and increased in their processing capacity exponentially. Unfortunately our interface with them has changed very little. ## Keyboards and Keymaps The keyboard, our primary means of input, it's offset keys and it's keymap are hangovers from the days of mechanical typewriters. The keymap, designed so that there was minimal chance of the hammers binding with the speed of a competent typist. The offset, due to the bars connecting the keys to the hammers. The qwerty keymap in common use today dates to 1878. Although neither of the above mechanical constrains exist and keymaps have been created, more optimally laid out for many languages. They have not been widely adopted. ## Mice Mice were designed in the 1970's as a secondary input device, but due to their intuitive nature, have come to rival or even take precedent over the keyboard for many tasks. Software has developed that exacerbates this trend by allowing draggable windows and bars. These software concessions make computers easier to access and more intuitive but then the majority of users never progress to learn how to effectively use the keyboard shortcuts. These are grave mistakes in my opinion and causing harm to many, as modern life, for an increasing number of people, requires many hours use of these devices. ## Injury to Health Documented conditions resulting from prolonged computer use: * RSI (Repetitive Strain Injury) to the arms and wrists. * Upper back and neck problems resulting from being in a fixed position for too long or with bad physical alignment. * Eye problems and headaches due to poor monitor refresh rates, brightness or just too long focusing on a fixed position. ## My Personal Journey A number of years after I began using computers for long hours I started to experience pain in my right wrist. This would persist for a while, even after I stopped using the computer. I'm fairly self aware and sought to address the problem on noticing it. Due to the pain being only in my right wrist and noticing it more when using the mouse for extended periods, I first tried switching to using the mouse left handed. Of course, that just moved the problem. I then went through a couple of years trying different trackball designs. Fortunately, by this time, I came to own an IBM thinkad with a Trackpoint mouse. While using it I noticed that I didn't suffer the same pain. After discovering there were USB versions of the Trackpoint keyboard, I bought one to use with my desktop computer. I continued looking for ways to further reduce my symptoms. On reading that a different keymap, more optimized for typing the English language (Dvorak), may reduce some fatigue, I decided to give it a try. After 6 weeks with a piece of paper above my keyboard, so I knew where the keys were, I achived parity with my previous typing ability. From this point on my speed and accuracy increased and I could use the computer for longer with little to no discomfort. ## Conclusion From the above you will notice that the mouse was the primary source of my problems. The wrist is not designed to be in a craned position for extended periods or for micromovements which cause strain to the tendons. Fingers are much better evolved for such fine movement. ### How can we improve our experience? * Ortholinear/Ergonomic Keyboards * Change keymap to one best optimised for your language * Pointing Stick Mouse built into the center of the keyboard * Learn keyboard shortcuts and make effective use of the keyboard I realise that changing keymap is a significant investment in time, but I would say that if you are brave or committed enough to make that investment, you will forever reap the rewards.

Stop-IT-waste

2021-10-01T00:22:23+02:00

# Stop IT waste What we buy determines what companies produce and in what quantity. If we buy their products in vast quantities we are telling them that they are producing what we want. Do you really want to buy devices which spy on you or have a pre-planned end of life, likely way before they physically expire? Non replaceable batteries, non upgradable components, time-limited software updates, specialist repair centres, use of low quality materials, all contribute toward an early end to the useful life for our devices. The term for it is planned-obsolescence. 1. The biggest drivers of this ever quickening turnover of hardware are mobile phone and tablet manufacturers. ## Personal experience I unfortunately own two android tablets, one is already obsolete and could only be rescued with a custom firmware. The other is heading the same way, even though it is a high specification tablet, there are no further firmware updates from the manufacturer and gradually apps are requiring newer and newer versions of android to run them. I remember when you bought a piece of hardware and it worked until it physically didn't. Any applications that were installed would continue to function as they did from the day you bought the device. Now it's quite possible for you to turn on your phone or tablet and discover an app crippled by an automatic update because the developer wants you to use something else or they decided to increase the minimum OS version requirement. This recently happened with the play app on my android phone and tablets. A basic feature that I bought the device with, mp3 playing, was suddenly crippled by an update. I couldn't revert back to the previous version, but I could uninstall all updates to return me to the factory installed app. This restored the function, but how many people would have just followed the notice to use youtube music or whatever? These practices are manipulative and fundamentally wrong. ## Reversing the trend My suggestion is for you to understand what you need and find suitable, preferably used, hardware to fulfill that need. There is only one way to signal to companies that you dislike their products, or the direction they are taking with them and that is to not buy them. Older computers are far more upgradeable than new ones and, with the correct choice of operating system, give you far more freedom over its use. For example, a core 2 duo with 2 - 4 Gb RAM will speedily run Linux or one of the BSD's, where as with windows 10 it will struggle. The majority of people can and should work effectively using Linux or one of the BSD's and readily available open source software. There are two reasons I suggest used hardware. The first, is that it sends no signal to the supply chain that there is more demand. The second, from an environmental standpoint, you are using something which has already been made and so making no further impact on the environment. To continue using items as long as you can is a big win for your bank balance and the environment; no new item has to be made and no old item has to be disposed of. ## Living by my principles ### Phone My daily carry is now a Nokia E71. 2. After thinking about what functionality I actually needed and how I'd been let down by various messaging apps (Fring, WhatsApp, Signal ...), I decided upon self hosting an XMPP server and reverting to SMS for contacts without XMPP. Finding an XMPP client for the Nokia took some time but it was worth it to be able to use my data for messaging again. Functions I use: * Calls * SMS * IM (XMPP) * Navigation * Alarm clock * Calendar * Calculator * mp3 player * Camera Possibly I will upgrade to an E72 or Blackberry Classic at some point. ### Computers I have two 4U ATX rackmount cases which I purchased around 15 years ago. The internals have been upgraded several times. One of them ran as a server for 12 years or more, the other has been a desktop. Upgrades have been purchased off Ebay, usually, with some new components like HDD's, drive caddies and PSUs from other sources. The most recent upgrade was to a used core 2 duo motherboard bundle which replaced an ailing dual celeron motherboard which had several leaking capacitors. 3. Foot notes: 1. [Planned Obsolescence](gopher://gopherpedia.com/0/Planned%20obsolescence) 2. [My Nokia E71](gopher://gopher.icu/0/phlog/Computing/Nokia-E71-in-2021.md) 3. I intend to replace the capacitors on the Abit BP6 board. Note: For full disclosure I do still have an android phone which I use at home, without a sim-card, for LINE contacts and Strava.

OCC-conclusion

2023-03-28T14:47:26+02:00

# Old Computer Challenge - conclusion I found that my habits and enjoyment changed considerably during the challenge due to a number of factors. ## Hardware My choice of hardware for the challenge was probably a bad one. Using old systems is one thing, but when you use old uncommon types that probably weren't particularly well supported when they were in common use, then you're in for double the pain. They are usually the first to be abandoned by developers. ## Software The most significant impediment to my being able to use the computer meaningfully was the software. Software rot, not due to the software itself changing or degrading but goal posts change due to external factors. My biggest bugbear was that I couldn't connect to modern websites using SSL/TLS due to the age of the libraries the software was built against. You might think, as I initially did, no problem I will just update the libraries and rebuild the software. Let me tell you, it's really not that simple, neither libressl or openssl would build on this system using the version of gcc and libc that I had. Compilation on such old hardware is also painfully slow, lots of wasted time trying to get stuff to build that ultimately failed. I don't have the knowledge or experience to understand or fix the errors reported by the compiler, that's on me. I found myself falling back to using the tmux session on my usual computer just to use lynx. I hoped I'd only have to do this to download the software libraries to fix the issues on the laptop but it was a futile exercise. If I had started this challenge again I should have at the outset tried to find a more recent operating system for the architecture. NetBSD would probably have been a good choice. NetBSD is a rare gem in that it still supports many of these old architectures which have been left behind by Linux. ## Final thoughts If you decide to do this challenge then do it with as modern OS as you can find. Don't expect to be able to connect to modern webservers with old versions of browsers, wget, curl etc. If you're lucky you will find a few that still have ftp, but non https webservers seem to be becoming more scarce. On the positive side I found myself using the computer a lot less and in a more focused way, because it was a bit of a chore and I didn't have easy access to things like youtube or podcasts. Getting into the habit of switching your computer off is a good thing. I leave my normal computer on 24/7 as it serves gopher, it's too convenient to just sit down and start wasting time. I have a VPS so I'm thinking I will move my gopher service there and also everything I currently run in tmux, so I can attach from any computer that has a terminal application. With a change to my backup strategy I should be able to switch this machine off when I'm not actually using it which will save wasting electricity too. I don't need to be connected to the internet all of the time.

Web2Atom

2022-01-20T14:58:52+01:00

# web2atom A simple utility for parsing data from a web page with multiple items and making it into an atom feed. Requirements: * Perl Get it: git clone git://gopher.icu/web2atom ## Commandline usage curl -s 'http://greatsite.com' | web2atom -p "great_site" ## Sfeed integration This utility was designed to work with sfeed[1]. To enable this modify ~/.sfeed/sfeedrc to look like the following: ``` feeds() { feed 'Ebay - ZX Spectrum' 'https://www.ebay.co.uk/sch/i.html?_nkw=zx+spectrum&_sop=10&rt=nc&LH_PrefLoc=1' 'ebay_watcher' #feed 'Another great site' 'http://greatsite.com' 'great_site' } parse() { if [ "$3" = "ebay_watcher" ]; then web2atom -p "$3" | sfeed #else if [ "$3" = "great_site" ]; then # web2atom -p "$3" | sfeed else sfeed "$3" fi } ``` The above is a working example. You will of course wish to tailor this for your own purposes. ## web2atom Inside the program you will find a list of profiles: ``` my %profiles = ( 'ebay_watcher' => { itm => '

.*?<\/div><\/div><\/li>', dmap => { ## Default Atom link => 'class=s-item__link href=(.*?)\?.*?>', title => '

(?:)?(.?)<\/h3>', published => 's-item__listingDate">(.?)<\/span><\/span>', content => '', ## Custom fields used in applyCustomFormating() price => '(.?)<\/span>', postage => '(.?) postage<\/span>', buyprice => '(.?)<\/span>', buyitnow => '(.?)<\/span>' } } ); ``` You must create a profile such as 'ebay_watcher' for each site you wish to parse data for. This entry must exist in the program itself and also in the feeds() and parse() functions in ~/.sfeed/sfeedrc as per the example . It is important that the 'itm' regular expression encompasses each item you wish to capture. note Setting the DEBUG constant to 1 will be useful while experimenting. Using firefox developer tools to view the html layout of a page and select the containers you're interested in is useful for creating the regular expressions. ## Default fields Treat the 'Default Atom' section as required fields for making a basic atom feed. The content regular expression is left blank as this will be populated later in the script with whatever is matched by the 'itm' regular expression. ## Custom fields The 'Custom fields' are totally optional and allow a great deal of flexibility. You can use the applyCustomFormatting() function to manipulate any of the captured fields. The below example shows how dates can be manipulated and how the title can be amended to include other data: ``` sub applyCustomFormatting { my ($src, @entry) = @_; foreach my $entry (@entry) { if ($src =~ /^ebay/) { # Fix and format date my ($date, $time) = split(/ /, $entry->{'published'}); $date .= qx#echo -n `date +"-%Y"`#; $entry->{'published'} = qx#echo -n `date --date="$date $time" +"%a, %d %b %Y %T %z"`#; # Append stuff to title $entry->{'title'} = "[BID $entry->{'price'} $entry->{'postage'}] - $entry->{'title'}"; if ($entry->{'buyitnow'}) { $entry->{'title'} = "[BUY $entry->{'buyprice'}]$entry->{'title'}"; } } } } ``` [1](https://codemadness.org/sfeed.html)

OCC-day2

2023-03-22T00:59:18+01:00

# Old Computer Challenge - day 2 ## Hardware issues The G3 Wallstreet is taking several attempts to start OS9 from where I have to launch the Linux boot loader. After several failed attempts with 'bus error' this morning, having to pull the battery and the power lead and try again, I decided to take it to pieces to reseat the ram and CPU. Having done this, I can now report that it's not made a bit of difference. It still refuses to start Linux at the first attempt. I also noticed that Linux isn't recognizing all of the ram. It only sees just over 100Mb when there is 512Mb installed. OSX sees the correct amount of memory but OS9 also only sees 128Mb. I tried adding 'mem=512M to the kernel boot line but it seems to have made no difference. ## Software issues The old software is becoming a bit of a problem in that the world has moved on. Even lynx and w3m text based browsers are refusing to work with some sites, including duckduckgo.com, as they must require a newer version of TLS. Configuration files from my existing system also don't work with the earlier versions of the same software (i3wm). This is proving to be a bit of a pain. I want to search for new libraries and software to download to fix the situation but I can't. Only recently I said "Hardware dies, software just gets forgotten.". However it would seem that if there are external factors, protocols, APIs, software can fail too. ## Thoughts The problem I have is that I'm using a 1998 32bit PPC mac. Most major distros have already left this machine to the digital dumpster, that's if they ever supported it at all. I have to decide whether to try to make do with this aged operating system and upgrade the software that's causing me problems or to try to find a newer OS to install. If the hardware wasn't misbehaving then this wouldn't feel like such a chore, but it really is starting to bother me. All of my daily computing is being done using this as a VT connected to the tmux session on my usual computer at the moment because I'm in this do I / don't I limbo.

VPS-vs-Home-hosting

2023-03-31T01:10:32+02:00

# VPS vs Home hosting When I was working commercially in IT I would always shy away from hosted services. You don't know exactly what the infrastructure is, how it's managed, whether the support are complete monkeys (until something goes wrong) or when they are going to put the price up. So many variables outside of your control... Then a couple of years ago I worked out how much it was costing in electricity to power my server and was faced with potential downtime switching to another ISP. ## Living the cloud dream I conceded, and got myself a VPS with A2 hosting. They had decent reviews and the price was very reasonable for an entry level VPS, which was all I needed. I had no issues for a year or more. Then there was a short outage and not much in the way of apology or explanation... Then my VPS became a 'legacy VPS' and they offered new VPS on some other platform that I could move to, but they could not migrate my existing setup. They also took this opportunity to put the price up, probably to encourage me to move so they could decommission whatever infrastructure my VPS was on. Then they had another outage which lasted best part of a day, again no apology or explanation, and nothing I could do but sit by and twiddle my thumbs until it came back online. So basically, now I'm in the position where I'd need to buy another VPS to run in parallel to the one I'm already paying for so I could migrate everything across to avoid an outage... ## Should I stay or should I go? The question now is do I get another VPS or go back to home hosting? I was looking at openbsd.amsterdam[1] at 64 euro per year. That's less than I'm paying now for twice the RAM and over twice the storage. The only issue I ever had with hosting from home is with email, I had to do some weird ssl tunnel before to get my mail out through my ISP's SMTP server. Then there was the maintenance of the physical server. The last one spewed its' capacitor guts all over the place, so it was in dire need of some TLC. But then I'm back in the position where if I switch ISP I'm going to have an outage, if the server dies it's on me to fix it and there's no off site copy of anything... [1](https://openbsd.amsterdam)

Compaq-Portable-386

2021-02-11T01:44:31+01:00

# Compaq Portable 386 I saved [1]this computer from being dumped in a skip many years ago after learning of its impending demise on a newsgroup. After I got it home I played with it for a week or so before putting it away. From memory it had DOS and windows 3.11 for workgroups installed. Everything worked, the motherboard battery was OK, so too was the original connor 40mb harddrive. The machine specs were: 6mb Ram 40mb harddrive 5.25 inch floppy ISA expansion box containing: * 10Mbit Network card * Full length soundblaster with a speaker held in place with blutak I removed the small speaker, blutak and soundblaster as it looked a bit of a mess. I still have the soundcard but not the speaker. ## The second inspiration Some years later I fired it up again, I can't remember why. Maybe just to see if it still worked. Unfortunately at this point both the CMOS battery and the harddrive had died. These systems have no BIOS to speak of and need configuring using a utility disk. Luckily I managed to find disk images in both 5.25 and 3.5 format which allowed me to set the time and drive parameters. I had a few old ide drives laying around and managed to find one that would work with the unusual disk numbering system. From memory I installed a 500mb hard drive, but it wasn't recognized as the full volume of the disk. It allowed me to install windows 3.11 again and I remember spending a lot of time trying to find and install a very small linux distro. I used it for a while but quickly tired of having to set the drive parameters using the floppy utility every cold boot. ## Jump forward to 2021 COVID lockdown, boredom, reminiscing about old computers and that they could still be useful. Lets see what I can do with the old girl! Planned fixes and upgrades: * CMOS battery, retrofit with CR2032 battery and carrier. * Install soundblaster card and hidden speakers. * Replace 5.25 floppy with more standard 3.5 (half height full width) * IDE to CF card adapter with CF card as main drive. * Install DOS, windows 3.11 and NetBSD 4.0.1 (last 80386) * [2]Cat5 to WIFI bridge (hidden inside the expansion box) The idea is to get the most out of her without making any structural modifications. No cutting of any of the plastic casings and remain externally original to the period. ## Update 21/01/2021 The CR2032 carrier and batteries arrived over the past couple of days so today, having all the parts, I set to replacing the CMOS battery. The original EAGLE PICHER LTC-16P-CO-F-S3 battery was fitted deep [3]inside the unit requiring removal of the harddrive and floppy assembly for access. This isn't ideal, so I decided to make a new one that was more easily accessible. I had spotted a handy vacant space down next to the psu that looked a perfect size for the battery carrier, with a little padding added to stop it sliding around. After making sure the assembly fitted snuggly in the [4]space, I cut the wires to the required length and refitted the original motherboard connector. The [5]replacement carrier wiring and battery, were checked with a multimeter before final assembly. ## Update 23/01/2021 The past two days are a blur as there have been many obstacles to overcome. I searched the entire house, including the loft space, but I couldn't find the 5.25" system disks anywhere. This meant I had to find the [6]DOS utility program again to create new ones. The 500mb harddrive was in fact only 120mb, so I was far more space constrained than I thought. This meant I had to settle on just installing NetBSD, as I was more likely to use it than either windows 3.11 or DOS. I could always boot DOS from floppy disk for games. ### Creating a utility disk It sounds like a trivial thing to do, apart from I no longer have any 5.25" disks. I managed to find a [7]3.5" disk image, but that needed to be a low density 720k variety which I also nolonger have. After much experimentation with various floppy drives and floppy disks I found that I could fool the drive using an old trick of putting tape over the right hand hole in the floppy and formatting it a particular way. Thankfully I had a USB floppy drive for such emergencies! To format the freshly taped floppy: linux: $ufiformat -f 720 DOS/Windows: format a: /T:80 /N:9 After that writing the disk image was straight forward: linux: $dd if=compaq-720.img of=/dev/ DOS/Windows: RAWRITE.EXE After practically a day of trying, this allowed me to run the utility to set the date; and then spend several more hours trying to guess the correct hard drive type from the 47 available options... ### Installing NetBSD With only a 120mb (unformatted) hard drive and 6mb of RAM I knew that this would be a challenge. I was not wrong! NetBSD 4.0.1 was the last to support 80386 processors. The minimum requirements were 4mb ram and 50mb hard disk space. Though there is a disclaimer of sorts "we do not know of anyone running with a system quite this minimal today.". I tried, the memory requirement was too high... Going back a little further, to 3.1.1, there is a boot-tiny.fs floppy image which had a GENERIC_TINY kernel. This booted and got me to the installation menu. I proceeded through the install, partitioned the hard disk (manually set the geometry because it was incorrect) and tried to configure the networking but I could not get the network to come up. Eventually I gave up and shut down the machine. Shutting down gave me opportunity to swap out the Kingston KNE2000 for a 3com 3c509. I tried installing again. This time I succeeded in configuring the network card and, with some modification of URL paths due to these old BSD's being archived, I could now install the basic package sets (base, kernel and etc). I finished up the install and rebooted. On booting up it just stopped, as if there was no bootloader installed. I remembered that I had manually altered the drive geometry because it was reading incorrectly from the BIOS. I thought maybe this was the problem so I rebooted to perform the install again, this time leaving the disk as recognized by the BIOS. Unfortunately this reduced the available drive size to around 70mb which was not enough to do the install. I decided to try selecting another drive type using the utility disk to see if I could find one that showed more space when running the installer. The result was much better, with closer to 100mb available without manually changing the disk geometry. I made a swap partition of 12mb (twice the ram) and used the rest for /. I proceeded to install the sets via the network and finished up the install. As I saw the NetBSD boot loader come up on screen and scroll through to a login prompt, I was so happy! My efforts had not been in vain and I had managed to get NetBSD installed on a 386 class computer with only 6mb of ram and in less than 100mb of hard disk space. [8]A success! There is still some tweaking required but I'm happy with how it's gone so far. Now I just need to get a dvorak keymap on there... ## Update 24/1/2020 I modified the network configuration to use DHCP. This required setting dhclient=YES in /etc/rc.conf and modifying the static ifconfig line in /etc/ifconfig.ep0 to read 'media 10baseT'. I discovered during the install that media type is not detected automatically on some older multi-media (BNC, AUI, 10baseT) network cards. The keymap has been a struggle. Aparently the GENERIC_TINY kernel doesn't have wsmux, which would ordinarily link to wskbd. The end fix was 'wsconsctl -f /dev/wskbd0 -k -w encoding=us.dvorak'. I have added the line to /etc/rc.conf for now to set it at boot. ksh needed a .kshrc to make arrow keys and command history work : set -o emacs HOME=/home/ export HOME HISTFILE=$HOME/.sh_history export HISTFILE While not essential, the above tweaks have made further configuration of the system much more efficient. ## Update 30/1/2021 The CF to IDE adaptor arrived today. I connected everything up and hoped that a CF to SD card adaptor with a 2gb SD card would do for testing. However, the system refuses to boot with the card fitted. I have ordered a 512mb CF card to see if it is the card or the adaptor that is the problem. ## Update 04/2/2021 The 512mb CF card arrived today, so I tested the CF adapter with the new card. After some time spent trying the various drive types I found one that seemed to give me close to the full drive capacity (type 41). After installing DOS it successfully booted from the CF card. Now that I was sure everything was going to work, I could start work to mount the CF adapter. The adapter came only with a steel bracket to mount it into the back of an ATX case. I figured that I could bend and drill the [9]bracket to fit into the drive rails without drilling any extra holes. This worked quite well until the point I realized the power connector was too close to the rail to fit. After some time thinking of a solution, it stuck me that I could unsolder the connector and place it on the back of the adapter which would make it face away from the rail. I was pleased with the [10]final assembly. I then cut one connector off an old HD ribbon cable for a best fit as the original was too short to reach the CF adapter. I then trial fitted and tested the adapter in situ. All was well, so I decided to install Windows 3.11 for workgroups. This was a little tricky as I had to modify the disk images to add support for the Compaq Plasma screen and CGA adapter. This makes the installation more straight forward though as you don't have to install drivers from alternative disks should it need to be reinstalled in future. To complete the retrofit I installed a soundblaster 1.5 that came with the machine and installed prince of persia and lemmings which would allow me to test the card. After making an IRQ jumper change everything worked great. I was quite impressed by the sound from this old card, despite nearly deafening myself when testing it with earphones! I reassembled the unit and gave it a final test by playing a few levels of lemmings with some powered speakers plugged in. The experience took me back to my teens when we had an Amstrad PC1512. ## Conclusion It was nice to have retrofitted the Compaq and have it working again without harming the integrity of the original system. I had ambitions when I started the project about maybe using it myself and installing WIFI and speakers in it. But really it belongs in a museum or with someone who will use it and appreciate it more. That is why I installed DOS and Windows 3.11 rather than NetBSD in the end. I set it up for a new owner rather than for myself as through the process I sadly realized that I would rarely, if ever, use it. 1. [Compaq Portable 386]( gopher://gopherpedia.com/0/Compaq%20Portable%20386 ) 2. [Cat 5 to WIFI bridge]( https://www.amazon.co.uk/Convert-Ethernet-Wireless-Repeater-Multi-Functional/dp/B07PMR673M ) 3. [Original battery location]( gopher://gopher.icu/I/phlog/Computing/images/old-battery.jpg ) 4. [Trial fitting new battery]( gopher://gopher.icu/I/phlog/Computing/images/trial-fit.jpg ) 5. [Finished assembly]( gopher://gopher.icu/I/phlog/Computing/images/new-battery.jpg ) 6. [Utility disk programs (3.5" & 5.25")]( gopher://gopher.icu/9/files/compaq-portable-386-utils.zip ) 7. [3.5" 720k floppy utility disk image]( gopher://gopher.icu/9/files/compaq386-cmosdiskimg.zip ) 8. [NetBSD Installed]( gopher://gopher.icu/I/phlog/Computing/images/install.jpg ) 9. [CF Card Adapter Bracket](gopher://gopher.icu/I/phlog/Computing/images/cf-bracket.jpg ) 10. [CF Assembly](gopher://gopher.icu/I/phlog/Computing/images/cf-assembly.jpg )

Gopher-Poll

2023-02-12T01:09:00+01:00

# Gopher Poll There was a gopher polling system[1] already, which I had used previously with no complaints. Unfortunately when I wanted to create a poll recently it was very slow, seemed to have issues creating the poll and the author didn't reply to my email, so I reverse engineered it. I had a good look at their system, took what I liked about it and changed what I didn't, which has resulted in something quite basic but functional[2]. To anyone who would like to test it or use it, please do and let me have your feedback. The code is available here[3]. [1](gopher://khzae.net/1/poll) [2](gopher://gopher.icu/1/poll) [3](git://gopher.icu/poll)

Nokia-E71-in-2021

2021-07-31T09:13:12+02:00

# Nokia E71 in 2021 ## Into the darkness Thinking back, the sole reason I stopped using my Nokia E71[1] was due to WhatsApp dropping support for the s60 platform. I had used WhatsApp for many happy years and my dependance on it, for keeping in contact with friends and family, was such that I switched platform to android to continue using it. ## WhatsApp bought by Facebook I have never liked/trusted facebook and so never created an account. News that WhatsApp was being aquired by facebook was a real hammerblow to me and I immediately expected the worst. When the updated terms of service, with an ultimatum to accept data sharing with facebook[2] or else, bomb dropped. I immediately sought an alternative and decided upon Signal. Unfortunately, in spite of my best efforts, my family and friends have continued to use WhatsApp... ## Android Paranoid I know it's spying on me, feeding data back to google at my expense, but I hadn't realized just how much until I looked at the data usage by application. Applications I didn't even use! If I were on an unlimited data plan then maybe I could turn a blind eye to it, but I'm not. Aside from this, there was a recent update to the Play app to cripple it. This was done for no other reason than to force users to youtube music. Fortunately I realized that by uninstalling updates for the app the functionality could be restored. Then the final straw, my contacts disappeared and gave some error about syncing. I have never enabled syncing of my contacts to google... ## Alternatives I looked at alternative options only to find that in modern times you really have only two, iOS or android. If that's not a monopoly I don't know what is. I didn't want to be part of either of those ecosystems so I started looking back in time. Firstly I considered a Blackberry Classic. I had missed the keyboard of the E71. The classic would provide newer hardware with a nice keyboard and a better camera, but I resisted the urge to buy the first reasonably priced one on ebay. In the meantime I decided to put my sim back in the E71. ## Fast forward It has been approximately four weeks since I reverted back to my trusty Nokia E71. This wasn't as much of a shock to me as it would be for many as I don't participate in any social media. The only apps I used with any regularity were Strava, WhatsApp, play (for music) and occasionally Maps for navigation. Since I had already switched to SMS and using Signal after closing my WhatsApp account. For the most part I haven't missed anything, other than the cheap messaging offered by an IM client that uses data rather than SMS which I currently pay 4p per message for. ## Opportunity After the WhatsApp disappointment and then reports of Signal offering digital payments[3] I had already began to think seriously about using an open IM protocol and self hosting. It didn't take long for me to decide upon XMPP, more commonly known as Jabber. Installing the prosody XMPP server[4] was fairly trivial, getting it to work with conversations/blabber on android however, was more tricky. As a client on the E71 I ended up installing 'slick'. It is fairly basic but at least allows me to exchange messages. I believe nimbuzz is more feature complete and will do voice calls too but I have yet to find a download. ## Software Finding software for s60 in the present day is a problem. Thankfully I still had backups of some useful apps: * BirdStep SmartConnect (Connection manager) * Opera Mini (web browsing) * Ovi Maps (navigation) I was actually quite fortunate to get Ovi Maps working again. Although I had the mapping application, I didn't have any of the map data saved and all the servers which hosted them have disappeared. Finding the map data[5] on archve.org and some helpful instructions[6] made my day. A nice bonus was to discover a gopher client[7], all be it written in java. ## Conclusion It's not perfect but then nothing is. I do however feel more confident in the phone. Knowing that the apps installed will continue to function as they always have, that my credit is not being wasted and that data about me is not being collected and used or sold for profit by google. [1](gopher://gopherpedia.com/0/Nokia%20E71) [2](https://www.zdnet.com/article/whatsapp-this-is-what-happens-if-you-dont-accept-our-new-privacy-terms) [3](https://decrypt.co/55422/signal-is-experimenting-with-stellar-based-cryptocurrency-report) [4](coming soon) [5](https://archive.org/download/Nokia_Maps_original_caches_from_Nokia_servers) [6](https://answers.microsoft.com/en-us/mobiledevices/forum/all/newest-nokia-ovi-maps-0125114-manual-download/7f0e8253-ea9b-4d64-9005-a7538dca468c) [7](https://felix.plesoianu.ro/mobile/pocket-gopher)

dallas-ds12887-cr2032-battery-mod

2023-03-02T21:46:24+01:00

# Dallas DS12887/DS12C887 CR2032 battery mod The modification is fairly simple but can be done in two ways. When I performed this mod I exposed only pin 20, for the positive feed, and cut a groove into the casing at pin 12, to expose a little of the pin shoulder to solder the negative lead to[1]. Most mods I've seen done have used pins 20 and 16, with the CR2032 battery holder I had it seemed tidier to use pin 12, YMMV. ## Pinout -------- MOT [|1 V 24|] Vcc NC [|2 23|] SQW NC [|3 22|] NC AD0 [|4 21|] RCLR AD1 [|5 20|] NC =========== + AD2 [|6 19|] IRQ AD3 [|7 18|] RESET AD4 [|8 17|] DS AD5 [|9 16|] NC =========== - AD6 [|10 15|] R/W AD7 [|11 14|] AS GND [|12 13|] CS -------- [1](gopher://gopher.icu/I/images/dallas-ds12887-cr2032-battery-mod.jpg)

OCC-day1

2023-03-21T00:48:24+01:00

# Old Computer Challenge - day 1 Where to begin, things haven't gotten off to a good start! Firstly, the G3 apple wallstreet powerbook I had decided to use for the challenge has been very unreliable to boot. It's OK once booted up, but if I have to restart the machine for some reason I'm getting bus errors. I didn't have this issue the last time I used it so maybe something is suffering a little from old age, maybe some capacitors somewhere? Amazingly the battery still charges! The version of Linux installed is Linux mint 9. It's based on Debian squeeze for which long term support ended back in 2016. All the repos have been archived and I had to mess around with the apt sources.list to get access to the binary packages to install / update software. I did briefly get annoyed to the point that I decided to attempt to use the dual pentium pro computer that I recently got running again. Unfortunately something caused a reboot and it has managed to corrupt the CF card I'm using as a HDD and ruin the NetBSD install that's on there. It also draws 120w, so it was back to the laptop for another attempt, after I'd calmed down from uninstalling network-manager which broke my wifi card configuration ... The version of i3wm in the repos is too old for the configuration from my current system to work, so it seemed like a good opportunity to give ratpoison a try. It's taking a bit of getting used to and I'm missing some kind of status bar at the top. I don't think there is one built into ratpoison so I may have to install some stand alone status bar. ## Software change summary Uninstalled lxde and associated packages. Openbox -> ratpoison lxterminal -> urxvt bash -> mksh Copied over my .mkshrc and .Xresources files.

Sabbatical-over-and-permacomputing

2023-01-25T16:31:23+01:00

# Sabbatical Over I'm not sure if anyone actually reads my phlog posts, or cares, but sometimes I think it's just good to put in writing the things that are whirring around in my mind, if for no other reason than to clarify my own thoughts. The end of last year was a bit of a rough time for me and I withdrew from everything, including being online and doing anything computer related, other than checking my email and RSS feeds maybe once a day. Maybe that's just what people should be doing rather than using computers in a non-deliberate and permanently connected way, as a distraction or for entertainment. Is there a difference? ## Permacomputing I've been reading quite a bit on the subject of permacomputing the last few days and have read a number of articles by Solderpunk[1], Viznut[2] and others. Orphans of Netscape[3] really resonated with me as I am one of those orphans. It's strange how there is a growing feeling among people that there is an inevitable shift coming in the availability of electronics and computing that we currently take for granted. No one is really sure what that landscape will look like, but there is a tangible nervousness with people trying to predict and prepare for what is to come. Whatever it is, the change will be far wider reaching than just our personal computing abilities and maybe computers as we use them today will cease to even matter... [1](gopher://zaibatsu.circumlunar.space/1/~solderpunk) [2](http://viznut.fi/texts-en/) [3](gopher://zaibatsu.circumlunar.space/0/%7esolderpunk/phlog/orphans-of-netscape.txt)

Quarry

2023-05-16T02:02:00+02:00

# Quarry (indexer, search interface and supporting tools) > A place, cavern, or pit where stones are dug from the earth, or > separated, as by blasting with gunpowder, from a large mass of rock. > Hunted or slaughtered game, or any object of eager pursuit. Quarry contains a number of components: 1. Crawler/indexer (quarry.pl) 2. Gopher search, front end to search index (search.dcgi) 3. Wrapper for quarry.pl to process pending host index requests (indexPending.pl) 4. Sitemap generator (generateSitemap.pl) 5. Host and selector maintenance (checkHosts.pl) Requirements: * Perl * curl * MariaDB/MySQL Try it: gopher://gopher.icu/1/quarry Get it: git clone git://gopher.icu/quarry ## 1. Crawler/indexer The indexer will by default visit every link on a gopher site and store the type, link-title, selector, hostname and port in the 'selectors' table. It will do this only for those types defined in HARVEST_TYPES. The robots.txt standard file format is supported and honoured. A bespoke sitemap file format is also supported and will be used to populate the database if found. There are a number of parameters which can be set at the top of the file to change it's behaviour: DEBUG (Default 1) Display verbose status messages. MAX_DEPTH (Default 5) Defines the maximum number of levels of recursion. IGNORE_ROBOTS (Default 0) Ignores robots.txt and any directives therein. IGNORE_SITEMAP (Default 0) Ignores the sitemap and instead indexes the site by recursion. CRAWL_DELAY (Default 2) Default delay in seconds between requests. This parameter is overridden by robots.txt if found. HARVEST_TYPES (Default '10Igs9') Defines the gopher types captured. TRAVERSE_DOMAINS (Default 0) *Best avoided* as sitemap and robots are only parsed for the start domain. It is better to index each host individually and use alternative means of host discovery. REINDEX (Default 0) Removes all selectors for host before re-indexing. Usage: quarry.pl some-gopher-domain.net The port can be optionally specified eg. some-gopher-domain.net:7070. ## 2. Gopher search This provides a front end to the index generated by quarry.pl. Features include: * General search * Image search * Sound search * Video search * Submit site to be indexed The current search function is basic and tries to match the search string against the selector or title fields and returns any that match. This will change once metadata is added and implemented in the search. ## 3. Wrapper This program simply looks in the 'pending' database table for hosts submitted to be indexed, via the gopher search front end, and passes them to quarry.pl to be indexed. Usage: indexPending.pl ## 4. Sitemap generator The sitemap generator uses data from the index generated by quarry.pl. The reasons for the sitemap are twofold: 1. Efficiency, downloading a single index file rather than crawling. 2. The format supports additional metadata: * Description * Categories * Keywords These extra metadata fields can be used to greatly enhance search results. Example of records: ``` Type: 1 Selector: /contact.dcgi Host: gopher.icu Port: 70 LinkName: Contact Description: My contact details Categories: Keywords: -------- Type: 1 Selector: /gutenberg Host: gopher.icu Port: 70 LinkName: Gutenberg (unofficial book and audio search interface) Description: Gopher search interface to the official Gutenberg book repository Categories: Books Keywords: Books -------- ``` Usage: generateSitemap.pl some-gopher-domain.net > sitemap.txt ## 5. Host and selector maintenance Only basic host checking has been implemented using 'hostcheck'. Currently the checkHosts.pl script checks each host twice in a 24 hour period. If the host fails two concurrent checks then it is flagged as inactive and selectors will not display in search results. If on a subsequent check the host has recovered then it is again flagged as active. Hostcheck: git clone git://gopher.icu/hostcheck ## 6. IndexNow IndexNow[1] is an easy way for website owners to instantly inform search engines about content changes on their website. It has been implemented in a basic way to allow submission of a single URI per request: curl -s 'gopher://gopher.icu/7/quarry/indexnow.dcgi?url=&key=' [1](https://www.indexnow.org/)

OCC-day5

2023-03-25T00:53:40+01:00

# Old Computer Challenge - day 5 I gave up trying to update software. For internet stuff I'm using weechat, sfeed, mutt and lynx via my tmux session on my usual machine. I have settled into using ratpoison and urxvt on this apple G3 laptop. The only thing I'm really missing is having a status bar to see battery life and time. I did briefly have a go at using i3status with dzen2 but no text appears in the bar that appears at the top of the screen, so I gave up on it. The time can be seen with 'C-t a' via ratpoison or 'C-b t' in tmux... ## Changing habits Today I have mostly been testing and scrubbing old SCSI hard drives on my Pentium Pro. I have a lot of old hardware which I don't need and requires testing. Seeing as it's now working again I'm making use of it for testing these old components. I have no plans to use SCSI in the future so I'm going to sell them on, hopefully to people who have a need for them to keep their old systems running. This afternoon I mowed the back lawn. I definitely find myself spending less time on the computer without a purpose. Due to not readily being able to waste time browsing Ebay or watching Yewtu.be videos, I guess I have been getting on with things which I've been neglecting. All this has made me realize that the computer I currently have running 24/7 could be switched off over night, and other times too. The only reason it stays on is to backup my VPS and serve gopher. Gopher can be moved onto my VPS and I can figure something out with the backup strategy. With it sat there running all the time there is always the temptation to just sit down and use it. If I had to wait for it to boot up then, unless I really had something to do, I wouldn't bother.

Simple-Computing

2023-01-26T00:03:42+01:00

# Simple Computing There appears to be emerging a renaissance in retro computers and retro computing technologies. My own earliest memories of computing are of playing games on my ZX Spectrum 48k and using BBC Model B's at school. In those days both hardware and software were much more simplistic and because of it, more accessible to hobbyists. Over 20 years have passed since I embarked on a career in I.T. and to be honest it was drudgery for the most part. Sometimes I was able to come up with some neat simple solution of my own devising but it was not often I had the freedom to craft something from the ground up. I also spent most of that time contributing to the commercial web for which I now have a lot of resentment. I much prefer tinkering with my old computers, they have character! Unlike the glossy screened featureless rectangles of today's masses. Not because of age, but because of the speed at which computing technology has moved on in my lifetime, I am a dinosaur of the digital age. To me this stuff isn't retro, it's what I grew up with, what I feel at home with and I will do my best to hold onto all that was good. ## Tools to create Every home computer in the 80's came with a built in programming environment and documentation allowing the new owner to extend the system by writing their own programs. A whole generation of game studios and legendary programmers emerged from this perfect storm. Unfortunately as commercial software took off the included tools allowing you to create your own steadily disappeared. In hind sight, if I were at the beginning of my career again, I would have learned C, Lua and POSIX shell. Forth has also recently come to my attention and intrigues me. Interestingly the Jupiter Ace[1] personal computer from the 80's came with a Forth interpreter as opposed to the usual basic interpreter of the time. Operating systems such as Collapse OS and Dusk OS are Forth based. We should be eternally grateful to those who have contributed to free and open source software. They have given us a lifeline from the tyranny of commercial operating systems and development software. I do believe however that much of it is far too complex... ## Plain text There are many text editors and command-line tools designed to work with plain text and this format is truly cross platform. I tend to write all my articles using markdown so that they can be converted easily to other formats which support text attributes like **bold**, _italic_ and headings etc. Although I rarely use anything other than headings and foot notes... I use vim for everything. Choose an editor and learn to use it well, this is your most important tool for working with text. ## Gopher Due to my resentment of the commercial web, and in some attempt to repent, I have become a gopher advocate and contributor. Through providing some small services (Gutenberg Search, Quarry Search Engine, Podcast Search) and writing content unique to gopher I hope to encourage its use and popularity. Gopher has not lost its community feel. Unlike the majority of the WWW, you don't have to look far to find content that has been created because it mattered to the author to share their knowledge or their opinion. ## A consideration To quote Solderpunk[2] "The dominant use of personal computers in the 21st century is the functional simulation of non-computers.". The emulation of analogue devices (clocks, calendars, typewriters, music players ...) and tasks. In response to that I'd simply like to say that it was only after the tools of creation were no longer distributed along with these devices and they lost the primary means of interacting with them in a meaningful way (a real keyboard) that they became devices of consumption. The clue is in the fact that the screen is the most significant part of them. To write software for android it would appear you need to download android studio which runs on what I would consider a real general purpose computer. They are not self contained multi purpose computers like those in the 80's. [1](gopher://gopherpedia.com/0/Jupiter%20Ace) [2](gopher://zaibatsu.circumlunar.space/0/%7esolderpunk/phlog/do-you-even-compute-bro.txt)

Unnamed-4U-computer

2023-03-09T17:49:19+01:00

# Unnamed 4U Computer Ok so only two people voted[1] on which of the two Dual Pentium Pro computers should be my next hardware project and they both voted for this one, so here it is! ## The hardware * Motherboard - ECS P6FX2A (Produced in 1996) * CPU(s) - 2 x Pentium Pro 200mhz (256k Cache) * RAM - 256mb 66mhz EDO (512mb max according to manual) * HDD - 8gb Lexar CF card connected via IDE to CF adaptor * Video - Matrox Millenium 16mb (dual vga output) PCI card * Network - Realtek 8139c 100mbit PCI card * I/O - 4 port USB 2.0 PCI card, IR and USB 1.1 board headers * SCSI - Diamond Fireport 40 PCI card with 9.1gb UW SCSI HDD * Case - Generic black 4U server case On first switching on the machine I was pleased to discover that there was in fact 256mb of RAM installed! Unfortunately the BIOS battery reported as being on its last legs, so that was to be my first job. I have written an article about the CR2032 battery mod[2] to the Dallas RTC chip, so I won't elaborate on that here. On reading the manual I discovered there were headers for both IR and 2 x USB 1.1. Also notable was that the board was capable of taking a maximum of 512mb of RAM. ## Grouch So this machine was unnamed but on thinking about it, the last host name I gave it was 'Grouch', and I'm beginning to remember why! This thing is quirky beyond all quirkiness. It won't boot from CD, gives some error 'AX = FF' after beginning to boot. I can't find a modern linux distro that will get beyond the first stages of trying to boot a kernel, all result in some kind of kernel panic, something about a trampoline??? The last operating system I had successfully run on it was Windows 2000 Pro and from memory I had to install it from floppy! Anyway, after much mucking about downloading and writing various linux iso images to a USB thumbdrive, which I had to initialise using plop[3] from a floppy as the system has no option to boot from USB, I finally got something to work, NetBSD[4] to the rescue! The USB install image worked right off the bat, no dramas at all. ## What to do with it? I have a lot of computer parts which are about the same vintage as this computer so for starters, I will use it as a test bed for those components. I can then have a clear out of things I no longer need and get them listed for sale so they can find new homes with people in need of them. ## Conclusion Its fans roar and its SCSI drive wails as it bursts into life. It's loud, ponderous and ungainly, a relic of the past. I can't help but feel satisfaction having extended the life of this digital dinosaur. As old and slow as it is, it could still manage 90% of my daily computing needs... Old-computer-challenge?[5] I'm feeling a lot of love for the NetBSD folks right now. [1](gopher://gopher.icu/1/poll?poll=1676405525) [2](gopher://gopher.icu/0/phlog/Computing/dallas-ds12887-cr2032-battery-mod.md) [3](https://www.plop.at/en/bootmanager/download.html) [4](https://www.netbsd.org/) [5](https://dataswamp.org/~solene/2021-07-07-old-computer-challenge.html)

vimb-web-browser

2023-02-08T03:24:27+01:00

# Vimb Web Browser I've been a firefox user longer than I can remember, but over recent years it has begun to annoy me with it's user hating interface changes and other negative aspects of the mozilla foundation. This has prompted me to look for an alternative. I'm also a long time vim user and for the past few years have used browser plugins to enable vim like key bindings to control the browser. The experience unfortunately, has always been somewhat lacking. vimb came to my attention due to it being webkit based and also having more integrated vim keybinding control. Building the software was quite straight forward following the instructions on the github page[1]. ## Requirements * gtk+-3.0 * webkit2gtk-4.1 * gst-libav, gst-plugins-good (optional, for media decoding among other things) From memory I think I had to install gstreamer1.0-libav on devuan to get youtube videos to play. ## Getting it to work as x-www-browser Initially my system kept opening firefox when I clicked on links in the terminal. To rectify this on devuan I tried to run update-alternatives --config x-www-browser Unfortunately it didn't detect the browser I'd installed from source as there was no reference to it in the dpkg alternatives which is located here: /var/lib/dpkg/alternatives/x-www-browser After adding an entry for vimb like the following: ``` /usr/local/bin/vimb 70 /usr/local/share/man/man1/vimb.1 ``` I could then successfully run update-alternatives and select vimb to be the new default. ## Conclusion It's taking a little getting used to but I'm starting to really like it. The fact that there are no tabs stops me from leaving lots of tabs open so it's actually changing the way I use the web browser, for the better in my opinion. [1](https://fanglingsu.github.io/vimb)

Gutenberg

2021-06-01T10:31:25+02:00

# Gutenberg Gutenberg is a fantastic free eBook resource with a truly enormous catalogue of over 60,000 out of copyright and public domain books. Through my gopher browsing I discovered that they still have a [1]gopher server which provides access to the text versions of all books, as well as some audio and video files. What they didn't appear to have was a nice way to search the catalogue using gopher. ## Gopher interface I had been thinking for a while how I could contribute to the gopher revival and this seemed like a logical union. Rather than providing a gopher interface to http resources, here was a gopher resource that was not being fully utilised. I set to work to rectify this situation. ## The index files In the directory listing there are several index files listed by year from 1996 in the format of GUTINDEX.. There is also a GUTINDEX .ALL and GUTINDEX.zip containing a consolidated catalogue listing. I chose this later file to parse and build a searchable index. Parsing the file was not without it's trials, with multibyte charact ers causing me some frustrations. However, perseverence allowed me to get what I needed from the file and build the index. Creating the file path URI's from the book numbers was fairly easy. An example of how to resolve the number to a directory was helpfully given in the README file. Example, book 1190. You split the numbers and each number would be a directory except the last. So /1/1/9/1190 where the book was single digit number you had to begin at zero. So book 9 would be /0/9. ## Gopher menu Initially I tried to link directly to the text files. Unfortunately the naming convention was not consistent and after guessing and programatically trying to verify the existance of the file. I decided to just link to the directory. This was a much cleaner solution. When I noticed there were [2]audio and video files as well I conceded that this was the best option. The end result is a [3]gopher menu which provides a search function that searches the catalogue for matching authors and titles. It also shows the five most recent additions to the catalogue and five most downloaded, with a link at the bottom of each section to list more. ## Update 01/06/2021 - Atom and RSS new book feeds added. After a request on IRC for a feed of newly added books, both [4]Atom and [5]RSS feeds have now been created. The Atom feed can be accessed directly from a link on the front page of the gutenberg interface. 1. [Gutenberg Gopher Server](gopher://dante.pglaf.org/1) 2. These can be searched for by searching 'Audio:' and 'Film:'. 3. [Gutenberg gopher interface](gopher://gopher.icu/1/gutenberg) 4. [Atom new book feed](gopher://gopher.icu/0/gutenberg/newbooks.atom.xml) 5. [RSS new book feed](gopher://gopher.icu/0/gutenberg/newbooks.rss.xml)

OCC-day3

2023-03-23T02:50:13+01:00

# Old Computer Challenge - day 3 I started the day with the usual unplug, battery out, wifi card out, half a dozen reboots to get linux booted. ## Software rot The day has been consumed by finding and downloading software dependencies and compiling them. This was made doubly difficult by the fact that the reason I'm doing this is because none of the installed web browsers will connect to large number of websites via HTTPS. It's a bit of a chicken and egg situation. Luckily some of the sites I needed to download software from still had ftp servers available. For the ones that didn't I resorted to downloading the source to my normal computer and then using scp to bring them across to the laptop. Needless to say old machines take a lot of time to build code from source, especially large packages like gnutls which have quite a lot of dependencies. I figured if I could update gnutls then I could at least build lynx and maybe a light graphical web browser so I can have some access to modern websites without cheating and using my normal machine in some way. By lunchtime I decided I'd had enough for a while and decided to go out and do some gardening[1]. ## The afternoon Later in the afternoon I got back to it and managed to assemble the remaining software dependencies to build lynx, either from source or binary packages. The first attempt at building lynx resulted in no SSL/TLS in spite of adding the necessary flags to the configure line. ./configure --with-ssl=/usr/local/lib --with-gnutls=/usr/local/lib --enable-gnutls-compat I figured something must have gone wrong with my build of gnutls; sure enough I hadn't enabled openssl compatibility, so I had to rebuild it again. I'm currently waiting for lynx to rebuild, fingers crossed it works this time ... [1](gopher://gopher.icu/0/phlog/Self-sufficiency/230322-Gardening-diary.md)

Old-computer-challenge

2023-03-16T13:37:53+01:00

# Old computer challenge I feel a little bit of a fraud being in #old-computer-challenge on libera.chat, having never actually done the old computer challenge[1]. So, I will be starting the challenge this coming Monday, 20-03-2023! ## Hardware I have decided to take the challenge on this occasion using an Apple G3 Wallstreet Powerbook[2]. Specifications as follows: * 300mhz G3 Single core processor * 512mb RAM * 60Gb HDD (spinning rust) * CDROM * 14" Screen * 1 mouse button (Surely this alone is a challenge!) ## Operating system Currently it has OSX 10.2 or 10.3 and an old version of Linux Mint. This may well change depending on my annoyance levels ... ## Initial thoughts I will try not to cheat and just use it as a terminal connected to my normal machines' tmux session. This should be an opportunity to test out some alternative software to that I normally use. I hope it still works ... [1](https://dataswamp.org/~solene/2021-07-07-old-computer-challenge.html) [2](https://everymac.com/systems/apple/powerbook_g3/specs/powerbook_g3_292.html)

Tablify

2021-10-06T14:13:05+02:00

For-Sale

2024-01-20T15:19:06+01:00

# For Sale ## Computers ### Compaq Portable 386 This computer[1] has 6Mb RAM, 512Mb HDD, 5 1/4" FDD and comes with the optional expansion box that can house two full size ISA cards. It has been retrofitted with a CR2032 battery for the CMOS and a CF card HDD without making any structural mods. A compatible working 117Mb mechanical HDD is included. It is pre-installed with DOS and windows 3.11 for workgroups. Two none original Compaq 5 1/4" system disks are also included. [1](gopher://gopher.icu/0/phlog/Computing/Compaq-Portable-386.md) Price: 250 GBP ### SGI Indigo 2 This computer has 256Mb RAM, 9.1Gb HDD, GR3 Elan graphics and a 20" CRT monitor. The keyboard, mouse and monitor are all the original granite finish items. It is currently running IRIX 6.5. There is also a 13w3 to VGA adapter to allow for use with a sync on green supporting VGA monitor. Price: ? ### HP N54L microserver Dual core AMD Turion 2.2Ghz processor, 4Gb of ECC RAM, HP 410p 512mb battery backed cache raid card and 2x 2Tb WD Red drives. Has been a great NAS and daily driver but I have downsized. Price: 250 GBP ### Dell sx620 USFF Dual core Pentium D 3.4Ghz, 4Gb RAM, 500Gb HDD and DVDRW. There is a 19" Dell monitor and logitech wireless keyboard with trackpad to go with. Price: 100 GBP ### Apple 12" Powerbook 1.5Ghz, 1.25Gb RAM, superdrive, bluetooth, wifi, 160Gb HDD. Comes with the original box, PSU, connectors and manuals. Battery is totally useless and won't charge. Price: ? ## Components ### SCSI HDD * IBM DNES-309170W 9.1Gb UW 68pin 7200rpm x2 * Seagate ST32430M 2.1Gb 50pin 5411rpm x2 * Seagate (CDC) 94191-15 668Mb 50 pin Full Height * Hot swap 5 1/4 bays for 68pin drives x2 ### SCSI Controller Cards * Diamond Fireport 40 - 50pin and 68pin PCI ## Mobile Phones ### Nokia 8110 (NHE-6BX) with refurbished battery The famous banana phone used in the Matrix (not the actual one) with a refurbished battery[2]. The phone is in good used condition and working order but I suspect locked to O2 and its derivative networks. I will in the next few days verify this. [2](gopher://gopher.icu/0/phlog/Computing/Nokia-BLJ-2-battery-refurb.md) Price: 90 GBP

Reduce-and-Simplify

2023-03-13T18:58:40+01:00

# Reduce and Simplify Over recent years I have gone through phases of consolidation and reduction in all areas of my life and that has inevitably included the hardware and software choices I have made. Consider with care that which you truly would not want to be without and know that it will continue to serve you tomorrow as it did today. ## Hardware considerations Try to buy hardware which adheres to standard form factors, like ATX, and uses common discrete components (not everything on a single board). This gives the largest scope for repairability and sourcing of replacement parts. At one point I had 3 computers running 24/7 eating up electricity. I now just have one computer running 24/7 which serves as my desktop pc, gopher server and NAS. Thinking about it I could probably switch it off over night if I moved the gopher server to my VPS and changed my backup strategy ... If items use batteries, buy ones that use standard cell batteries (AA , AAA, C, D) that are easily replaceable or at least ones that are removable and fairly easy to source (Nokia BP-4L / BL-5C). ## Software ### Operating System I migrated to Devuan linux from Debian because I didn't like the fact that systemd was becoming much more far reaching than just an init system. It was ceasing to be something that was an option and was fast becoming an ever more inextricable part of many linux distributions. This was a simple transition for me as Devuan is Debian based. The primary reason I started using Debian all those years ago was due to its advanced package management system. Void linux was also on my radar when I was looking to move from Debian and systemd as it provides init system options and there is also a Musl library based version. I decided against it as it has its own package management system and I didn't really want to have to learn a new one. I also like and would consider using any of the BSD family of OS, having spent a little time using all of them over the years. pkgin now seems to offer all the functionality of Debians apt package manager. ### Window manager Xfce -> i3wm - I started using a tiling window manager to maximize usable screen real estate and minimise distractions, along with CLI application replacements for any GUI ones that I happened to still be using. I have recently been toying with the idea of giving ratpoison a try. My typical desktop[0]. ### Terminal xterm -> urxvt - Urxvt is a lot smaller in size and has some nice extras. It allows you to configure URI recognition so you can click on URI's in any terminal application to open them. Using this in combination with a plumber program, which decides on the correct application to open them with, makes for a very slick integrated feel: I click on youtube links and it opens them with mpv, I click on gopher links and it opens them in lynx, http and it opens vimb etc. Kudos to __20h__ for the plumber utilities. I use tmux in combination with the terminal application to essentially give myself virtual workspaces with different applications so I can easily switch between them. Also it means I can detach tmux and restart my windowmanager if there's an issue and reattach without having to re-open everything. ### Editor Vim - I use vim for everything and use vim keys to control other applications wherever possible (vimb, ksh, lynx, ...). For years I used vim commercially and without a clue as to its power, not even scratching the surface of its features. A text editor is a vital part of your tool kit, invest the time and learn to use it well. ### Email Thunderbird -> Mutt - It was a coin flip between alpine and mutt. I found a few good tutorials on setting up the mutt config file so after all the time invested in setting it up how I like it, I've stuck with it. ### RSS Newsboat -> sfeed_curses - Brilliant RSS feed reader, kudos to Evil_Bob for such a great piece of software. I use it every day and it saves me from the tyranny of the WWW by bringing content I'm interested in to me. ### WWW + Gopher Firefox -> Lynx/vimb - I find myself using Lynx more and more as it's a great text based browser which also supports Gopher out of the box. All modern graphical web browsers are huge and bloated and there's no way around it. Firefox, I started to have some ethical as well as usability issues with. Fortunately after a bit of searching, keeping in mind my love of vim keys, I discovered a webkit based browser called vimb. For the most part is has been nice to use and has changed the way I browse the web. I no longer leave lots of inactive tabs open. ### IRC Xchat -> weechat ### Music Rhythmbox -> mocp ### Video VLC -> mpv + yt-dlp ## Programming languages Perl was the first programming language I learned, and my general go-to language, but it has quite a large footprint. There are worse options, but I know now that there also lighter alternatives. I will certainly be looking at using Lua in place of Perl for future projects. Although in the meantime I have found myself using shell a lot more for small jobs. I should also put some real effort into improving my C programming as many projects are written using it in the open source world. Forth looks an intriguing, minimalist, low level language, maybe I will dabble with it a bit too at some point. ## Conclusion I hope the above will give some inspiration for those looking to reduce and simplify their hardware and software footprint. If you manage to have a more focused computing experience or get a few more years service out of your existing system from anything I've written then, to me, it's been worthwhile. If you have any suggestions or alternatives to any of the above software, that you think I should take a look at, then please get in touch. [0](gopher://gopher.icu/I/files/desktop.png)

OCC-day4

2023-03-24T11:37:00+01:00

# Old Computer Challenge - day 4 I've learned my lesson, to avoid spending the first 10 - 15 minutes of the day trying to get linux to boot, I'm now leaving the laptop switched on. The sleep and resume function seems to work well and the battery also amazingly still holds about an hour of run time. ## lynx It still didn't build with gnutls and ssl. I must have tried 10 times now, with various flags and paths, and all I'm doing successfully is cooking my balls! This thing gets hot! ## Distractions I decided to fix the NetBSD install on the Pentium Pro machine and reset the BIOS settings back to optimal in the hope that whatever caused it to reboot last time and hose the installation won't happen again. I've installed xfree and a few other bits and pieces. It's giving me grief over not having termcap for urxvt but that's the least of my worries at the moment. At least the version of lynx on there can access HTTPS websites! I refitted the SCSI card to the Pentium Pro while I was at it and decided to start scrubbing the old SCSI disks I have laying around so I can get rid of them. The first 9.1Gb drive is done and the second one is most of the way through as I write this. Hopefully it will be done soon because I want to go to bed and this thing is pretty loud. From down stairs it sounds like there's a small truck running in my bedroom. On the positive side it has been running all afternoon and not missed a beat. ## lynx pt2 I figured maybe the openssl libs which are installed are too old so I downloaded libressl and tried to compile that. That too fails to build so it's not looking promising I'll get lynx to build any time soon. I tried asking on #lynx on libera.chat but no one answered... ## Thoughts Software dependency rats nests! There's nothing standard about anything, not even the compression tools used to compress the files for download. How on earth package maintainers deal with all this and stay sane, I'll never know!

openbsd.amsterdam-vps

2023-05-15T00:42:47+02:00

# OpenBSD.Amsterdam VPS So I did it, purchased another VPS and decided to commit to a year with openbsd.amsterdam. So far I have not been disappointed, the VM was set up during a holiday! I get regular emails about planned updates or service additions, nothing trying to sell me anything. I'm very happy so far. ## Server Migration If you have read my previous post you would know that I already had a Linux VPS with A2 hosting. I expected there to be some snags moving to openbsd but some of them I really wasn't expecting. Many of my shell scripts would not work without significant modifications, even though I used mksh on Linux. Many of the problems were due to GNU extensions to base utilities like sed, awk, date that I'd inadvertantly used. I got so fed up at one point that I installed gawk just to have inline file modification. Firstly I moved my gopher hole and all associated software, there was a lot more than I realised.. I then moved my git repositories and set up git. This transferred everything I was hosting from home which has allowed me to start shutting down my desktop computer at night and hopefully save a little on the electricity bill. Next I moved my static business website and domain, while leaving the email where it was. I then copied across the remaining websites, which included a MediaWiki and Wordpress sites. After that I dumped the databases and copied those across too. Getting the websites to work was a little more involved than I thought. Webservers on OpenBSD (openhttpd and nginx) both run inside a chroot. Due to the way partitions are assigned on the VM /var doesn't have much space so I wanted to put the websites on the home partition. In order to achieve this I had to change quite a few paths , including that of the php-fpm.sock. It proved to be a bit of a faff but I got there in the end. Also the database wouldn't connect on localhost but for some reason worked on 127.0.0.1. After I got the first site working I, one by one, changed the DNS to point the domain names to the IP address of the new server. During this process I set up nsd, which is the OpenBSD domain name server. I had been meaning to host my own DNS again after I had some issues transferring my business domain. On transferring the domain from my previous registrar I had to wait and didn't know when it would happen exactly. When the domain did finally transfer none of the zone information was carried across, resulting in a short outage. It wasn't the first time I'd had problems like this and I didn't want that happening again so now I'm hosting it again myself. All that remains now is to move my email, which I have been procrastinating over because the last time I did it, things didn't go so well...

kindle-offline

2023-05-26T19:32:14+02:00

# Kindle Offline ## Privacy I highly recommend switching off 'whispersync for books' and amending the 'Privacy' settings of your device. Better still don't connect it to your WIFI... ## Connect to Linux Connect your kindle to your computer with a USB cable; the device should display 'USB Drive Mode'. Identify the kindle drive using 'dmesg': [ 4802.467725] scsi 7:0:0:0: Direct-Access Kindle Internal [ 4802.468745] sd 7:0:0:0: Attached scsi generic sg2 type 0 [ 4802.473757] sd 7:0:0:0: [sdb] 6385664 512-byte logical blocks: [ 4802.579692] sd 7:0:0:0: [sdb] Write Protect is off [ 4802.579705] sd 7:0:0:0: [sdb] Mode Sense: 0f 00 00 00 [ 4802.689915] sd 7:0:0:0: [sdb] Write cache: enabled, read [ 4802.924022] sdb: sdb1 [ 4803.169675] sd 7:0:0:0: [sdb] Attached SCSI removable disk Find the UUID of the device: # blkid /dev/sdb1 /dev/sdb1: LABEL_FATBOOT="Kindle" LABEL="Kindle" UUID="####-####" BLOCK_SIZE="512" TYPE="vfat" Add UUID to '/etc/fstab': # Kindle UUID='####-####' /mnt/kindle auto user,noauto,nodev 0 0 Make directory on the filesystem to mount it to: # mkdir /mnt/kindle Mount the kindle: $ mount /mnt/kindle You can now copy books in .txt or .mobi format directly to the device; simply copy your files to '/mnt/kindle/documents'. Unmount the device before disconnecting the USB cable: $ umount /mnt/kindle ## Get free books from Gutenberg I have written a small utility[1] for getting books of the correct format onto your kindle once mounted using the above instructions. There is also a daily updated RSS feed[2] of newly added books to the Gutenberg catalogue which can provide the book number for use with the above utility. Just take the last number from the resource URI: gopher://gopher.pglaf.org/1/6/7/6/5/67650 And execute the following command, after installing the utility. $ getGutenbergBook 67650 [1](git://gopher.icu/gutenbergToKindle) [2](gopher://gopher.icu/0/gutenberg/newbooks.atom.xml)

Usenet

2023-06-14T12:52:09+02:00

# Usenet My earliest memories of using the internet date to around the mid 90's. Back then I was on dial-up, like the majority of people who had any personal internet access. The multitude of giant internet and social media companies were in their infancy and if you wanted to communicate with people regarding some subject, you would either do it through IRC or Usenet news groups. ## 90's Dial-up era In the dial-up era, every ISP used to provide news servers along with email servers for their customers. There were newsgroups on practically every subject and they were very active. I got my first IT job through a posting on a newsgroup, I also bought and aquired computers and other hardware through them. Then Yahoo auctions and shortly after Ebay came to the fore as places to buy and sell. Myspace became a thing and everyone started making a personal website... I don't really know how or why, but newsgroups dropped off my radar entirely as everything seemed to be made accessible via the WWW. ## Escaping 2020's WWW monoculture Today everything is provided via the WWW or some app, everything has converged to that advert, popup, javascript infested mess and mass of sponsored content designed to keep you engaged and feeding your data back to the mega-corporations. In the beginning there were numerous competing and complementary protocols. NNTP (network news transfer protocol), gopher, WAIS, HTTP, FTP, finger, telnet, SMTP, POP, IMAP... Few are those who actually use them natively any more. I had already decided I wanted out. I was already partly out due to my adoption of RSS, gopher and return to IRC, but there had to be a way of further removing myself. What was it that I did before the boom of the WWW? Newsgroups, that's where I used to engage with people on things that I was interested in, or when I needed to know something that I couldn't find an answer to... ## Usenet - I'm BACK! Firstly no ISP's today seem to provide news servers. If you do manage to find a free provider, they are few and far between and will probably not offer the binary groups as this requires so much more storage and bandwidth. www.eternal-september.org are one of the few remaining free usenet access providers. Kudos to them for providing free accounts for people to access this resource. After registering for an account I looked for a CLI application to subscribe to and read the groups. slrn is a common one I believe, but on initial usage I couldn't get to grips with it. I ended up installing tin, which somehow seemed more intuitive. I managed to subscribe to a few groups, some that I remember from years past and a few new ones that more reflect my current interests. I decided to send a message to eternal-september.test, to make sure everything was as it should be, entitled 'Usenet - I'm BACK!'. To my great surprise I got replies! I had the feeling of stumbling into some hermitage where the inhabitants were only too happy to have visitors! My initial happiness unfortunately was all too quickly replaced by a feeling of melancholy at the dates of some of the last posts. Some were years previous, others had maybe one or two posts, that weren't spam, within the last 12 months. I found myself replying to old posts or posts I normally wouldn't just to generate a bit of traffic and let people know there were others still out there too. Usenet was a great resource and can be again. Get yourselves a free account[1] and reclaim it before it is lost forever. [1](gopher://bitreich.org/1/lawn/c/newsgroups.gph)

IRCNow

2023-06-16T23:58:01+02:00

# IRCNow ## Of the Users, By the Users, For the Users I first became aware of IRCNow through a talk Aaron Lin (jrmu) gave during Libreplanet 2022[1]. I was enticed by both the will to create an open user run network and also the opportunity to learn more about OpenBSD. At the time I was in the process of looking for another VPS provider and considering moving to openbsd.amsterdam. It appeared to be a great match for me. ## Joining the team To cut a long story short I joined their IRC network and introduced myself to jrmu. We talked for a while and he set up a VPS for me to do their minutemin training programme. It took me a while as I had other things going on in the background but eventually I completed the majority of the training and teams were asked if they wanted someone to join their team. I ended up being invited to join team nastycode[2]. ## The network There are 13 servers that should have a team assigned to each. Why 13 you may ask? It was a little while before I learned that it was because there were 13 founding states of the USA... Having 13 servers is actually a bit of a hindrance both technically, as you obviously need enough staff to man them, and financially. My objections to this have thus far fell upon deaf ears and partly due to lack of human resources we have issues with something or other on an almost daily basis. The servers are a federation, each team has control of their own server, runs their own choice of services and the only thing that really joins us all together is IRC. I get along with most of the other admins but seem to have gained the AKA of IronJ for my spirited arguments with the ones I don't. ## Team server (nastycode.com) When I joined, the disk usage of the server I am now sysadmin of was 105% on /home and 90% on /var. I didn't actually know it was possible to go beyond 100%. The disk was so full I couldn't actually delete large files and had to whittle away at things that looked safe to delete until the machine started to respond more normally. I found a lot of problems: expired SSL certificates, wrong permissions on directories that were part of the mail systems filesystem, backups in peoples home directories, which were also backed up when /home was backed up. Frankly it was a mess and not what I'd signed up for. I expected to be a junior and have someone to help and guide me with looking after the server but as it turned out I was pretty much left to my own devices for 2 days before another sysadmin appeared. To be fair though jrmu has been really good in helping out when I've needed assistance and between us we got most of the critical outstanding issues resolved pretty quickly. ## Progress I've ironed out a lot of the niggles on the server. A lot of the routine stuff is now automated to keep log file sizes sensible and certificates updated. Things have gotten better this past week. We have a new team member, gtlsgamr. He is admin of tildevarsh.in[3], a pubnix system in India. He was initially coerced by jrmu into helping out with creating a new wiki for IRCNow but he seemed keen to join a team. Based on the work I'd seen him do up to that point, I was more than happy to have him join ours. We had already talked quite a bit when discussing the wiki so he was aware of my stance on minimal software and complexity. This has resulted in us building our own team wiki and website more in line with our own philosophy; a git repository of markdown files and a static site generator (mkdocs[4]). ## Join us! There's never a dull moment over at IRCNow.org, if you're an admin anyway... Come and join us to chat[5] or maybe even do the Minutemin Bootcamp[6] training; learn how to admin OpenBSD and join one of the teams! [1](https://media.libreplanet.org/u/libreplanet/m/ircnow-of-the-users-by-the-users-for-the-users/) [2](https://nastycode.com) [3](https://tildevarsh.in) [4](https://www.mkdocs.org) [5](ircs://irc.nastycode.com:6697) [6](https://wiki.ircnow.org/index.php?n=Minutemin.Bootcamp)

Ode-to-Gopher

2023-06-19T20:07:56+02:00

# Ode to Gopher Like many others who's phlogs I have discovered in the gophersphere. I am frustrated with what the web has become. I remember the days before broadband and rich media, where text was king and the signal-to-noise ratio was much greater. Let us return to those naive days of discovery and wonderment! Long live gopher!

IRCNow-update

2023-06-20T23:50:18+02:00

# IRCNow update ## All good things must come to an end After several weeks/months of fighting with uninformed hobby hackers I finally decided it was a futile effort and to call it a day; I resigned. I could go on about all the problems at IRCNow and who was to blame for what, but it would do no good whatever. The idea is sound, if it wasn't I wouldn't have put the time and effort into it that I did for as long as I did. But there are always those in any group that feel threatened and band together to obstruct the 'movers'[1]. [1](gopher://morena.rip/020230618)

Thinkpads

2023-06-21T13:45:18+02:00

# Thinkpads My personal history with IBM Thinkpads goes back a long way. I am a serial Thinkpad owner, my first Thinkpad was a 760ED, followed a while after by a 560 (I loved this machine) that got stolen and then the R31 that replaced it. ## My previous Thinkpads I loved the modularity of the 760ED and the ultra-portability of the 560. Unfortunately the R31, although far superior in performance, was a bit of a letdown. After the R31 I lost my way and in 2005 ended up purchasing a 12" apple powerbook... ## The return Of late my tastes have been somewhat retro with regard to hardware and my love of Thinkpads has never gone away, I even use a Thinkpad USB Trackpoint keyboard on my desktop computers! I had already decided in my mind that I would keep an eye out for a vintage, yet usable, example if one should present itself. Candidates were x200 or x61. The x61 reminded me so much of my second love, the 560. With its square format screen it had the perfect retro aesthetic and was of enough age that it still appeared to have that genuine IBM quality that has all but been lost since their aquisition by Lenovo. With a dual core processor and maximum ram capacity of 8gb I was pretty sure it could easily meet my meagre mobile computational needs. I had happened to mention my Thinkpad preference in passing on #gopherproject and as sometimes rarely happens, the stars aligned. One of the regulars, lumidify, happened to have an x61s that he didn't need, which required a battery and harddrive but was otherwise in good condition and working order. He offered to send me the laptop and I couldn't refuse. A couple of days later he notified me that the package was in the post and with that I identified a fast 120Gb SSD (Samsung 850 Evo) and bought the cheapest second hand one I could find on Ebay. ## Package arrives I had been waiting in anticipation for the arrival of the laptop and soon enough there was a knock at the door. I was initially surprised that I wasn't asked to pay some import tax on the package before it was handed over. But once the door was closed my attention was fully on what was inside the box. It had been well packaged, as I'd hoped, and the charger brick and cable kept separate from the laptop so nothing got bashed together in transit. Taking a first look at it, I was amazed how small it seemed, was this really a 12" screen? It was so slim and light, it reminded me so much of my 560 that it was love at first sight... ## First inspection On opening the screen I could see that it was a bit dirty, especially around the bezel so I got out the foam cleaner, some clean cloths and a fine boar bristle brush. I first cleaned the power brick, as I'd been told it was a bit smelly as it had been owned by a smoker. I gave the cable and plastic housing a thorough clean with the foam cleaner and cloth. I turned my attention to the x61s itself, first cleaning the exterior and then opening it up to thoroughly clean the screen, bezel and then keyboard. And there it was, a few scars telling of it's road warrior life but you could tell that it had been cared for and not abused by its former keeper. ## Spirit of the machine Machines like this deserve care and respect. Treated well they will serve you for decades, if you don't get caught up in stupid fads of the web and corporate software... OpenBSD and I will be its new companions, hopefully for many years into the future.

Advent-of-Computing

2023-06-23T17:34:55+02:00

# Advent of Computing I don't usually go out of my way to promote something unless it is truly worthy of note, but recently I have been really enjoying this podcast[1]. It is well produced and full to the brim with geeky informative goodness; a must for any retro computing enthusiast! [1](https://adventofcomputing.libsyn.com/rss)

The-Ascetic

2023-06-24T21:33:18+02:00

## The beginner Adopter, adherent, consumer and proliferator of all new things. The novice is easily lead to new fads in all things tech. ## The intermediate More cautious, the process of discernment begins to arise. Merits and weaknesses of all things are weighed. ## The ascetic Only that which is needed is created, the superfluous is discarded. Nothing but that which serves remains.

OCC-July-2023-Prep

2023-06-28T14:27:32+02:00

# OCC July 2023 - Preparation The last time I did this challenge I did it with practically no preparation. I just made sure the machine I had intended to use actually switched on the previous day. On that occasion[1] I quickly realized that this lack of preparation had been a huge mistake and so was keen to avoid the same suffering I had during the last challenge. ## My champion I intend to do this challenge using a somewhat revised configuration of 'Grouch'[2]. This is a 1996 dual pentium pro 200Mhz system with 256Mb of RAM and an 8Gb CF card as the HDD; I will likely add a soundblaster ISA card. The operating system is currently a fairly bare install of NetBSD. This will change throughout the challenge as I try to construct a workable environment for my computing needs around the limitations of the hardware. ## What's the point? The point of my participation is two fold: 1) Proving actual old hardware can still provide the majority of every day needs. 2) Maybe the hardware limitations will force me to some software choices and solutions that I hadn't before considered. [1](gopher://gopher.icu/0/phlog/Computing/OCC-conclusion.md) [2](gopher://gopher.icu:70/0/phlog/Computing/Unnamed-4U-computer.md)

Smart-terminal-not-client

2023-07-13T11:26:11+02:00

# Smart Terminal - not client For some time I have been using what is referred to as a plumber program. This is a program which decides which application to open a file with based on its URI type and file extension. For example gopher:// links open in my chosen gopher client, http:// links in my chosen web browser, image files in my chosen image viewer etc... Used in combination with a terminal application which supports URI matching, like URXVT, it has effectively given me the ability to open links in any terminal application. This also has the added benefit of making ordinary text files into interactive documents[2]. ## Putting it together 1) Download and install the plumber[1] program. 2) For URXVT add the following to your ~.Xresources file: URxvt.perl-ext-common: default,matcher URxvt.matcher.button: 2 URxvt.matcher.launcher: p $0 URxvt.matcher.pattern.1: \b(\w+\:\/\/\S+)\b URxvt.matcher.launcher.1: p $1 URxvt.keysym.Mod1-Shift-M: perl:matcher:list Line 4 above adds matching of any URI to URXVT. Line 5 passes the matched URI in $1 to p which is a symlink to the plumb script. Button 2 is middle click and can be used to click on links to cause them to open using the plumber; alternatively you can use alt+shift+m to display a numeric menu of links that can be selected by pressing the respective number of the link. ## Update 12/07/2023 I had written a few patches for the plumber in the past but after this last install I wasn't particularly happy with calls to unknown programs being hidden away in the opener files. In an attempt to improve this I wrote a patch to require specifying the more common programs as environment variables. .xsession: ... # plumber programs export XTERM='urxvtc' export PLUMB_IMAGE='feh -F --auto-rotate' export PLUMB_MEDIA='mpv' export PLUMB_GOPHER='gopher' export PLUMB_TXTGOPHER='gopher' export PLUMB_PDF='xpdf' export PLUMB_FILEMANAGER='ranger' export PLUMB_WEB='lynx' export PLUMB_TXTWEB='lynx' ... [1](git://r-36.net/plumber) [2](gopher://gopher.icu/I/images/interactive.png)

grouch-occ-prep

2023-07-12T23:54:57+02:00

# Grouch OCC prep ## Hardware Grouch[1] with the following additional hardware: IDE * 8Gb CF card with IDE to CF adapter (found to be faulty so added PCI SCSI controller and SCSI HDD) PCI * Matrox Millenium G450 GFX (replaced with Nvidia GeForce FX 5200) * Realtek 8139C generic 10/100NIC (replaced with Netgear 10/100) * Belkin USB 2.0 card (USB header on motherboard doesn't work) * SCSI controller (added after CF card found to be faulty) ISA * Soundblaster AWE 64 ISA sound card (removed as caused hang in X) ## Software * vim * ratpoison * urxvt (missing terminfo on NetBSD) * gopher (crashes in X on OpenBSD) * lynx (SSL error: unable to get local issuer cert .. NetBSD) * mpg123 (never used this befoe but moc looks ugly out of the box) * git * mutt (compiled without sidebar on NetBSD) * xnotify ## Software configuration and optimisations ### Fixed size fonts This may sound bizarre but it was actually the biggest single usability improvement I made. When I installed X it seemed usable but after I copied my .Xresources file over from my other system, which included xft:dejavu sans mono with antialias things really did start to chug and I didn't realize the issue straight away so persevered with it. That was until I was chatting on IRC and the annoyance of watching each line shuffling up the screen like a caterpillar became too much to bear! .Xresources URxvt.termName: rxvt URxvt.cursorUnderline: false URxvt.cursorBlink: true URxvt.font: x:9x15,xft:FontAwesome:style=Regular URxvt.boltFont: x:9x15B,xft:FontAwesome:style=Bold URxvt.scrollBar: false ! map ctrl+shift+c/ctrl+shift+v to copy/paste URxvt.keysym.Shift-Control-V: eval:paste_clipboard URxvt.keysym.Shift-Control-C: eval:selection_to_clipboard URxvt.perl-ext-common: default,matcher URxvt.matcher.button: 2 URxvt.matcher.launcher: pj $0 URxvt.matcher.pattern.1: \b(gophers?\:\/\/\S+)\b URxvt.matcher.launcher.1: pj $1 URxvt.keysym.Mod1-Shift-M: perl:matcher:list ### urxvt in daemon mode (urxvtd + urxvtc) This improves the speed of launching new terminal instances maybe at the expense of stability, not that I have noticed any issues. To facilitate this I started the daemon in .xinitrc and then exited it again with a hook in .ratpoisonrc. As long as you :quit ratpoison properly then it will kill urxvtd. .xsession: xrdb ~/.Xresources setxkbmap -option caps:escape urxvtd & ratpoison I was a bit annoyed that I had to specify the keymap again after setting it up during the NetBSD install. On OpenBSD it works correctly for both terminal and x. Also terminfo was not configured for urxvt and I had weird colour issues that I didn't have on OpenBSD , even after following the recommendation to 'export TERM=rxvt'. .ratpoisonrc: bind c exec /usr/pkg/bin/urxvtc addhook quit exec pkill urxvtd set fgcolor orange set bgcolor black set barinpadding 1 set padding 0 20 0 0 set msgwait 0 set startupmessage 0 set winliststyle row ## Conclusion There are annoyances with both NetBSD and OpenBSD. OpenBSD: 1) Really slow bootup due to relinking (fixed by modifying /var/db/kernel.SHA456 - make last hash char a comment) 2) X doesn't launch properly and I have to switch between VT 1 and VT 5 each time I launch it to be able to see the DM. (fixed by changing GFX card) 3) Sometimes the DM shows corrupted fonts and the cursor as a big square fuzzy block (HW conflict?). (fixed by changing GFX card) 4) Annoyingly displays console log on DM and then again when you start the WM. (fixed by editing /etc/X11/xenodm/Xsetup_0) NetBSD: 1) set hostname but on reboot back to localhost 2) Similar corruption of fonts sometimes when launching X 3) missing certs in lynx 4) mutt compiled without sidebar 5) No proper termcap for urxvt 6) After specifying dvorak keymap on install, I had to add it to .xinitrc as for some reason it went back to qwerty Currently I'm not sure whether to go back to OpenBSD. I was expecting performance of NetBSD to be a lot better but it's actually not... Just the lengthy post boot relinking annoyance and the having to switch back and forth between VT's to get X to display are the main concerns to switching back. ## Update 8/7/2023 I reverted back to OpenBSD and have managed to resolve the previously mentioned issues. Fixes have been listed above next to the issue. Unfortunately I discovered another issue which was that the ISA sound card would hard hang the system if I tried to use it while in X. For some reason it was fine using it from the terminal but once X was started it would cause it to hang. I tried removing all the other cards and also using different GFX cards but to no avail. Unfortunately this meant that I now needed a free PCI card for sound, as I didn't have another ISA card, and I didn't have one due to having to us a PCI SCSI controller because of the faulty CF card... ## Update 9/7/2023 I had to make a call between going the week without sound or buying a replacement CF card so that I could revert to IDE and free up the PCI slot currently occupied by the SCSI controller. CF cards are expensive for the capacity so I was reluctant to buy one but it actually seemed like the only sensible option. I had already made the decision to sell off my SCSI drives and cards. Continuing to use them was just adding to my procrastination over selling them. ## Update 10/7/2023 It was late in the day when the replacement CF card arrived but arrive it did so I set to work. First I wiped the old CF card and a small IDE drive I'd used to check if there was a problem with the IDE controller. I fitted the new CF card and attempted to start up the machine but it was a no-go. The bios seemed to recognize the card as an ATA device but at that point it hung and would proceed no further. I tried reconnecting it and restarting the machine a few times to be sure but it wasn't going to work. I conceded that I would have to do the challenge using the SCSI subsystem so I reconfigured it to boot from that. While in the process I remembered I had an external USB sound card. It certainly wouldn't hurt any to give it a try. I was surprised that OpenBSD recognized the card on boot and on testing I was able to play sound. This meant I wouldn't have to do the challenge without audio; some small relief. [1](gopher://gopher.icu/0/phlog/Computing/Unnamed-4U-computer.md)

Old-Computer-Challenge-July-2023

2023-07-17T01:26:42+02:00

# Old Computer Challenge July 2023 ## Day 1 Day one was a bit of a write off due to me not wanting to start investing time in setting up the software on the machine as I thought I would have to start from scratch once the CF card arrived. I had some basic settings that I'd worked out during my preparation[1] for the challenge that would give me a good starting point. Toward the end of day one the CF card arrived but sadly the computer didn't like it and would hang after the point of identifying it as an ATA device. To cut a long story short I ended up back on the SCSI subsystem and with an external USB sound card, so at least I had audio for the week. ## Day 2 Tuesday was mostly filled with configuring software. I had some annoying display issues with urxvt that I coudln't seem to resolve and also the plumber[2], which I like so much, wouldn't open any of my links. I figured this was down to the environment not being correct. I was a bit disappoited by the end of the day and felt I hadn't made much progress. Highlight of the day was probably discovering I could take screenshots without installing additional software, after installing mpv anyway... ffmpeg -y -f x11grab -s 1920x1200 -i $DISPLAY -vframes 1 -vf scale=1280:-1 screenshot.png ## Day 3 Today was a better day. I finally resolved the display issues with urxvt and the problems with the plumber. Both issues were down to the environment not being set correctly. The soultion in the end was adding '. ~/.kshrc' to .xsession so that the environment variables etc were properly set before the window manager was started. I think the display issues were down to no defined locale settings. .kshrc: export PATH=$PATH:$HOME/bin export LC_ALL="en_US.UTF-8" export LC_CTYPE="en_US.UTF-8" export LANG="en_US.UTF-8" HISTFILE=$HOME/.ksh_history set -o vi alias S='function ddg { lynx "https://lite.duckduckgo.com/lite?q=$*"; }; ddg' alias W='function wik { gopher -p lookup -T 7 -i "$*" gopherpedia.com; }; wik' export EDITOR=vim .xsession: . ~/.kshrc # plumber programs export XTERM='urxvtc' export PLUMB_IMAGE='feh -F --auto-rotate' export PLUMB_MEDIA='mpv' export PLUMB_GOPHER='gopher' export PLUMB_TXTGOPHER='gopher' export PLUMB_PDF='xpdf' export PLUMB_FILEMANAGER='' export PLUMB_WEB='lynx' export PLUMB_TXTWEB='lynx' xrdb -merge ~/.Xresources xsetroot -solid black setxkbmap -option caps:escape urxvtd -q -f -o ratpoison I managed to get a bit of work done today too. Several updates to my phlog, including: * Addition of index file for /phlog directory to list all posts * Removal of '-' and '.md' from post titles * Update to 'grouch occ prep'[1] post to cover final hardware changes * This post In addition to this I made some code changes to the plumbing program and subimitted a patch for those changes. That is the reason for the PLUMB_\* environment variables above. I pulled several of the hard coded program names out of the individual opener files and made environment variables for them to be defined. Outstanding issues: * Problem with mpv/ratpoison stretching the display of small format videos * Seemingly no hw acceleration in mpv with this nvidia fx5200 Both of the above issues make watching video practically impossible. ## Day 4 I spent most of the morning wrestling with mpv and eventually came up with some settings that worked acceptably with ratpoison; to stop it from stretching the videos and to completely black out the rest of the window. .confif/mpv/mpv.conf: --video-unscaled --no-keepaspect-window --ytdl-format="best*[width=640]" The last line is just to tell yt-dlp which video format to fetch. 640 wide is about all this thing can manage without hardware acceleration which I have, so far, been unable fix. Having checked the Xorg.0.log I can see that the card is using the 'nv' driver for my NVIDIA GeForce FX 5200. But there is a notice that explains my lack of hardware acceleration: AIGLX: Screen 0 is not DRI2 capable Today I considered installing my Usenet and RSS clients, but I resisted the impulse to do it. There is a certain tranquillity that exists as a side effect of my current hardware limitations. I spent some time reading other peoples experiences on gopher, listened to a podcast and got my K&R C book out of my book case with the aim of beginning at the beginning, for about the third time, and progressing my way through the exercises in the book. hello.c: #include int main() { printf("Hello, world! "); } ## Day 5 Well, I started out with the intention of continuing with my C exercises but somehow that was put asside quite quickly. Quarry[3] became the focus of the day as I was very aware that I'd neglected it for some time. I hadn't re-indexed any of the sites I had already in there for months and I had never gone back and removed dead links. That was something I had intended to fix some time ago but hadn't got around to. I managed to crash my webserver while running the re-indexing script, that caused me a bit of a panic as I had to login to the vm host for the first time and use vmctrl. It was all a bit daunting as I couldn't email for help as my email is hosted on the server doh! This is the first day that I've just used the computer without making any tweaks to configuration or adding any software. ## Day 6 and 7 Not much to report really. I spent day 6 continuing my work on quarry and most of day 7 I was out. When I did return home I continued my work on quarry and read about the experiences of other people taking the challenge, on gopher of course! ## Conclusion I didn't get off to the best start in spite of putting about a week into preparing the system I intended to use for the challenge. I was surprised that OpenBSD performs as well as it does on this system. I fully expected NetBSD to be my only option but OpenBSD was, for me anyway, a much more polished experience. Unfortunately I didn't manage to get hardware acceleration working or I may well have been able to prove youtube to be watchable at low resolutions. It was nice to have a fresh setup for the first time in several years to experiment with. I had used ratpoison for my last challenge earlier in the year but hadn't put much effort into configuring it last time. In the end I found myself using tmux as my window manager and ratpoison windows as virtual desktops; one for local and one for my vps. The temptation to install tin (for usenet) and sfeed (for rss/atom feeds) was resisted so I was better able to stay focused on doing something useful. The older system has just enough latency to make you consider more consciously what you are about to do. When all is said and done, I have quite enjoyed the challenge on this occasion. I have proved that I can be productive using this 27 year old computer and I don't feel that I have been lacking for anything. [1](gopher://gopher.icu/0/phlog/Computing/grouch-occ-prep.md) [2](gopher://gopher.icu/0/phlog/Computing/Smart-terminal-not-client.md) [3](gopher://gopher.icu/0/phlog/Computing/Quarry.md)

Quarry-gopher-search-engine

2023-07-20T00:36:25+02:00

# Quarry gopher search engine Quarry[1] began life somewhere in the middle of 2021. After spending several weeks working on it and getting it reasonably functional, something undoubtedly happened which took my focus away from it. Since that time I hadn't really invested any time in it. I hadn't re-indexed any of the sites I had already in there for months and I had never gone back to remove any dead links. That was something I had intended to fix some time ago but hadn't got around to. To cut a long story short, towards the end of last week the subject of Quarry came up and I found some time and motivation to put into it once more. ## Re-indexing To begin with I ran the indexer against the existing hosts that I had in the database. This was unbearably slow and took the best part of 3 days to complete. When it was complete I found several problems with the data that it collected. I had some rogue strings (\xC2\x80\xC2\x98yo...) breaking my database inserts. I thought this may have been cured by setting up the perl script to use utf8 for input and output but it doesn't appear to have made much difference, if any. A bit of modification to various sleep times that I'd put in various parts of the indexing code however, made a big improvement to the speed. I had made some small database change to set an update date stamp whenever a selector was added / updated so that I could spot ones that hadn't been updated after an indexing run and remove them. This gets rid of stale links collected from menus but currently there is no checking before adding selectors to the index. ## Search results There was a lot of junk collected during the initial re-indexing. In order to deal with this I modified the rudimentary filtering that I'd created first time around and extended it with more filters as I manually trawled through the database and also from what I was seeing in search results. I also wasn't happy with what was being returned to searches, the results didn't seem particularly relevant in spite of my using a fulltext search index. It turns out that I had also indexed the selector itself, probably not the best idea. Having removed that index and matching only on the title, the search result relevancy improved quite a lot. This also had the added benefit of reducing the number of results returned. ## Future indexing I created a script to take all of the hosts and iterate through them, checking they are active before running the indexer against them and finally removing any selectors that weren't updated, meaning they were no longer present on the site. I ran this a couple of hours ago and the total indexing process took 6 hours and 44 mins, a bit better than the 3 days it took previously. Now that this is working I will put it on a cron-job and run it once a week. If you have a site and don't want to wait, there is an API feature you can use to request spidering or you can add your URL through the search interface[2]. ## To Do * Fix the rogue characters breaking my inserts * Extend filters so they can be restricted by host * Verify selectors before including in index [1](gopher://gopher.icu/0/phlog/Computing/Quarry.md) [2](gopher://gopher.icu/1/quarry)

Darkmode-Saves-Power

2023-07-26T13:09:53+02:00

# Darkmode Saves Power I had a thought recently after changing the background colour of my terminals and desktop wallpaper to black due to getting eye fatigue. It got to the point where my left eye would weep and due to this it got very sore. Often, late in the evenings, I could feel my eyeball going into spasm; not a pleasant experience! Because it made such a difference to my eyes and was noticeably less bright I pondered how much less energy was being used due to the black background, so I did a little test. I connected a power meter to the 24" IPS screen and monitored the wattage. Firstly with a white background (10.1w) and again with a black background (8.3w); a saving of 1.8w. Now 1.8w doesn't seem like much. However, like anything that's switched on for long periods of time, it all adds up! For mobile devices with limited portable power, this can also make a significant difference to run time...

4G-Challenge

2023-07-26T17:49:57+02:00

# 4G Challenge I've been with my current ISP, even though it has changed names several times, for over 20 years. Unfortunately over recent years they have continually increased the prices and more recently it is even in the terms that they can increase the price during the contract! We are now paying a discounted price of 32Gbp per month. ## Playing the game A little under 18 months ago, just as our contract was about to end, we had done some looking around and there were cheaper alternatives. None of them could match the outright speed of my current connection, but I don't believe I needed that amount of speed in the first place. We were all set to leave, but they offered to match the price of my proposed replacement provider and let me keep my existing service. It seemed like a win, no need to change anything and I get the price match. Unfortunately I didn't consider the in-contract price rises which means my cushy 25Gbp a month deal turned into the current 32Gbp about half way through the contract. ## Falling cost of mobile broadband I'm not falling for the same thing again this year so I am determined to leave. Over recent years a lot of investment has been made into 4G and 5G technology and I am always seeing adverts regarding its speed, which now largely rivals some of the more basic fixed fibre offerings. Pricing has also fallen dramatically and unlimited data SIMs are available for around 20Gbp a month. I happened to spot a good deal of 16Gbp a month for the first 12 months and so got myself a SIM card, along with a 4G modem, to see if it could be a viable replacement for my fixed broadband; before my current contract ends. ## 4G Setup I'm using a Proroute H685[1] 4G industrial router, which I purchased second hand off Ebay, and a SIM with unlimited data on the 3 network. I've been using this setup for the last couple of days and apart from the occasional bit of lag, it is performing very well. Even my Eurosport streaming subscription seems to be working without any problem at all. Speed[2] of the connection seems to fluctuate but even with the standard antennas I get around 20mbits down and 10mbits up. If things work out then maybe I'll invest in an outdoor antenna to try to improve the signal a little. [1](https://proroute.co.uk/proroute-4g-routers/proroute-h685-4g-lte-router/) [2](gopher://gopher.icu/I/images/4g-speedtest.png)

Computer-Upgrade

2023-08-10T02:49:52+02:00

# Computer Upgrade I have been using a HP N54L Microserver as my daily driver for about 8 years. It's a small square unit that doesn't consume much power (56w idle), has been installed with 4Gb of RAM, 2x 2Tb WD Red drives and a HP p410 battery backed raid card with 512mb cache. It also has a low profile soundblaster sound card to provide sound. The machine has served me well, but as time has gone on the fact that it is a non-standard form factor has been an increasing concern to me. If you have read any of my posts about hardware preferences[1] you would know that I advocate for standard ATX form factor computers because replacement parts and upgrades are far more readily available. I have had a more powerful computer sat idle for some time in the form of a 4U rack server case which holds standard ATX components. After it stopped being my webserver, when I moved hosting to a VPS 4 years ago, I bought a 3Ghz core 2 duo motherboard bundle with 4Gb RAM off Ebay quite cheaply to replace the Abit BP6 which had several leaking capacitors. Since puting it together I've added a further 4Gb of RAM and then done nothing more with it. Now seemed like a good time to put that more powerful system into service, if I could transfer the raid system across from the HP and ensure it would run reliably. ## Xeon modification I had been aware that it was possible to fit a socket 771 Xeon processor to some of the 755 boards, with a little modification. Some months ago, after a little research, I had decided to purchase an E5450 Quad Core Xeon to attempt this modification. This is a Quad Core 3Ghz CPU with 12Mb of cache. The 771 socket is a little different to the 755 so requires you to buy a little sticker to reverse two of the pins on the CPU, you also need to file 2 new grooves into the CPU, or remove the locating tags from the CPU slot on the motherboard. I didn't fancy potentially ruining the motherboard attempting to cut off the locating tags so I took to filing the CPU instead. It took a few trial fittings and repeated filings to get it to fit nicely but after a little work it fitted well. The sticker was a little fiddly to get perfectly aligned with the pins but also went without a hitch. I refitted the CPU fan after cleaning both the top of the CPU and the heatsink face with IPA and applying a thin film of arctic silver thermal compound. Firing up the PC was a bit nerve racking as I had never tried this before. The computer bleeped and then powered off, I thought maybe something was up with it but pressed the power again just in case. I saw some warning about the bios battery, it had obviously died since I had last used the machine but it proceeded through the boot sequence and appeared to show the correct CPU values. There was an old install of Debian on the HDD which it proceeded to boot. Checking 'cat /proc/cpuinfo' showed that there was indeed a Quad Core Xeon fitted[2]. ## Waiting on a mini SAS cable So the machine appears to work, but in order to move the raid subsystem from the other machine I need a mini SAS cable. I have one on order and it should be here in the next couple of days. Further updates will follow. ## Update 01/08/2023 The SAS cable finally arrived this morning. I hurriedly pulled apart the HP microserver to get out the hard drives and raid card that I needed to transplant into the Xeon. However, there was just enough doubt in my mind not to proceed with cutting the cable ties securing the cache battery to the chassis. Assembly of the components was simple as I had already refitted the hot swap SATA drive bays and the raid card was just a PCI-E card slot secured with a single screw. The cable was a little tricky as there wasn't much space at the back of the drive bays. I powered on the machine, but there was no display. I didn't remember having any problem with HDMI previously but just to rule it out I tried the DVI port, with the same result. I decided to check the BIOS to see if there was anything. I noticed that there was a setting for video which was set to auto detect, you could select internal or PCI-E. I wondered if it had gotten confused by the raid card so I set it manually to onboard. Rebooting again resulted in the same black screen. I pulled out the raid card and tried again. This time I got the boot screen. This proved there was no issue with the display so I refitted the raid card and disconnected the cable in case this was a source of the problem. Still there was no display. For whatever reason the system doesn't seem like the HP p410 raid card. I fitted my old PCI 3ware raid card, which has no cache or battery back up, to see if this worked; and it did. I continued to fit two 200gb drives that were previously in the same chassis when it was my former web and email server. This made it into a usable system but not the one I had hoped for. There was nothing more I could do but clean up the inside of the HP microserver and refit the parts I'd taken out. Thankfully everything worked just as it had before when I came to power it on again. I was honestly quite disappointed at this point. I'd spent the morning stripping down a perfectly serviceable computer to transplant the best disk subsystem I own into the most powerful computer that I own, only for it to fail to work. I've since installed OpenBSD on the Xeon machine and done a little configuration. I'm actually writing this post on it, but I really don't know what I will do with it in the long term... ## Update 04/08/2023 To cut an ever longer story short. I bought an Adaptec ASR-5405Z 512mb cache SAS raid controller off Ebay for 11Gbp. I was already quite heavily invested in this thing so I thought, what's another 11Gbp? I hoped that this generic raid card would not have the same conflict as the HP one and that life would be good again, but I also braced myself for the fact that it too may not work for the same reason as the HP raid card. Well, I fitted the card and connected it up with the SAS cable I had bought to connect the drives to the HP controller and ... Blank screen, same as the HP controller. I wasn't about to give up and thought there has to be something I can do. So I removed the raid controller and set about downloading BIOS updates. After several attempts I managed to make a bootable CD with the necessary software on it. Everything looked good but it just seemed to hang at the point of writing the first part of the update, I left it for quite some time as I didn't want to interrupt it if it was actually writing data but after some 20 minutes of it sitting there with the power light blinking and the screen stuck at 'writing intel me...' I decided to pull the plug. Next I managed to put a bios update file on a USB stick which could be read if it was vfat with the file on and accessed from the built in 'f7' bios update menu. It stuck at exactly the same place in the same way. I was getting pretty desperate by this point so I pulled out the xeon and put the core 2 duo back in to see if this made any difference; It didn't. Last resort, RTFM. So I started digging through the motherboard manual and stumbled upon the following passage: ``` When a non-video PCI Express add-in card is installed in the PCI Express x16 connector on the Desktop Board, the HDMI port is disabled. ``` "You've got to be kidding!!" so I found myself a dvi -> vga adapter and put the raid card back in and tried again. It worked! What a palava! I'd been pulling my hair out all because the HDMI connector is disabled if you fit a card in the pci-e x16 slot. Just to be sure I connected the SAS cable and configured the drives. Sure enough everything worked as it should and the day, and my sanity, has been saved! ## Update 09/08/2023 The saga continues. After my last update it seemed that I was home free and only had to install my OS. Well unfortunately that wasn't the end of it. I tried to install OpenBSD but it turns out there is no driver support for that card (adaptec ASR-5405Z). I was back to square one... At this point everything came into question, should I install FreeBSD which does have a driver? Should I just go back to Linux which also has a driver, or should I just go back to my old computer with OpenBSD as I could see that the p410 raid card was supported? I tried the p410 raid card in the Xeon for a second time and it seemed to work at first, I could see it initializing the raid. But just as I thought I was out of the woods, and it should have started reading from the disk, the screen went a light grey and everything appeared to hang. I tried a couple more times just to be sure it wasn't some weird timing issue but to no avail. I actually conceded defeat at this point and installed OpenBSD to the HP microserver but the pci-e Sound Blaster sound card wouldn't work... Now more determined to stick with the plan, I put the 3ware raid card back in the Xeon and began to use it to make sure it was going to be viable as my daily driver. Everything appeared to work OK and it was fairly snappy, apart from some noticeable lag of the disk subsystem. As I saw it I had 3 options: 1) Buy a supported SAS raid card, knowing I would lose HDMI 2) A hybrid system - SSD for OS and PCI raid with HDD for storage 3) SSD only - ditch the raid, small SSD for OS and second bigger one for storage In an attempt to speed the system up, future proof it and reduce its' energy consumption; I have decided upon option 3. This will consist of one 120Gb SSD main drive and a 500Gb SSD for storage. i I can't realistically ever see myself filling 2Tb. I managed to back up everything I have onto the 500Gb SSD with quite a bit of space to spare and I know there is a lot that could be deleted... [1](gopher://gopher.icu/0/phlog/Computing/Stop-IT-waste.md) [2](gopher://gopher.icu/I/images/xeon.png)

Low-power-computing

2023-08-17T14:57:42+02:00

# Low power computing If you have been reading along with my adventures lately, you will be aware that I had been upgrading one of my old ATX machines[1] to be my new daily driver and things hadn't been going too well. Having gotten the machine to a useable state I put it on the power meter to find it was drawing 68w at idle, which jumped to around 100w when scrolling around in firefox-esr. I really wanted this machine to be my forever system that I would simply upgrade and repair as time went by, but 68w ticking over is quite a lot. The HP microserver comparatively was 56w. Could I save 10w by removing the spinning rust and raid card? Possibly... ## Mighty Atom I can't remember the exact motivation for my next move but I had an old asus eeebox b202 which had previously been my mom's. After she had finished with it I had installed FreeBSD and used it for about 8 months, after upgrading the memory in it to 2Gb. From memory it wasn't exactly fast, as you'd expect from a 1.6ghz Intel Atom, but it was perfectly usable. I had earlier in the day checked the specs and had noted the maximum video output was 1600x1200, which was a bit of a disappointment as I have a nice 1920x1200 24" monitor. Undeterred I fetched it from its resting place, connected it to my monitor & keyboard and fired it up. It booted into FreeBSD and to my amazement it was at 1920x1200 when I checked xrandr. After a futile attempt to update FreeBSD (no idea why I even tried) I made an i386 miniroot USB stick to install OpenBSD. I have done this quite a few times recently so it didn't take long, especially now my all my dotfiles are in a git repo and easily deployed with stow. It was disappointing to discover there was no firefox-esr in i386 OpenBSD and I had to seek an alternative for those times I really must use the web. Ungoogled-chromium seemed like a viable option so I installed it to give it a go. It was just a bit too slow and clunky to be tolerable. On checking the power meter I noted that the system was drawing only 17w... Encouraged by the fact that everything worked so well out of the box with OpenBSD and that the system only consumed 17w, I pondered what I could do to make it a potential daily driver. The next day I would take delivery of a 120Gb SSD meant for the xeon... When the SSD arrived I had already decided to try it out in the Atom. Removal of two screws allows the HDD tray to slide out revealing four more screws to remove the drive. I went through the same installation procedure as the previous day and in under 20 minutes I was back at the point where I could again try out ungoogled-chromium. The SSD had pepped the machine up quite a bit, booting was noticeably more speedy and everything seemed that bit more responsive. Chromium certainly wasn't fast, but it was at least now tolerable. Just enough sluggishness not to make you want to do too much on the web; it was perfect! I used it for the rest of the evening. This morning I took it apart and made a new BIOS battery[2], as the old one had died forcing me to reset the date every time I unplugged it. With the new SSD, and a second one plugged in via USB, it draws just under 16w and does everything I need[3]. [1](gopher://gopher.icu/0/phlog/Computing/Computer-Upgrade.md) [2](gopher://gopher.icu/I/images/asus2.jpg) [3](gopher://gopher.icu/I/images/asus3.jpg)

Do-more-with-less

2023-09-01T02:01:30+02:00

# Do more with less We have been convinced that more and newer is always better, faster, more convenient, which if you're in the business of selling products is exactly what you want. But in the interests of personal development and sustainability I have actively been pursuing the opposite. In the modern world we are suffocated with choice, from toothpaste to programming languages, and constantly bombarded by technology vying for our attention. The more stuff you have the greater the burden of ownership and the less time and attention you can give to any one thing. ## Computer downgrade I recently downgraded[1] the computer I use daily to one that is adequate for my needs and consumes just 16w of power[2]. This wasn't really a planned downgrade at the outset but rather the result of some experimentation and long held philosophy regarding sustainability and minimalism. I have been using the system for over a week now and it is adequate for how I use it. Though I realize such a system would be impractical for most due to their software choices. Web browsing is possible using ungoogled-chromium, the sites that I have needed to use, for online banking or ebay, have all been quite slow but usable. This is more of a feature than a problem as it makes browsing the web tedious and so it is kept to a minimum. Mostly my local system is used much like a terminal, with many of the internet applications (Email, IRC, Usenet and RSS feed reader) running on my VPS. My thinking here was that, if I needed access to the internet to use them, they might as well exist on my VPS which reduces the amount of configuration and software I need on my local system(s). ## Software footprint For a while I have been actively trying to use lighter weight software for my daily tasks and have migrated from heavy GUI applications like Thunderbird, VLC and Rhythmbox to lighter TUI alternatives. I typically use: * mutt - mail client * tin - Usenet news reader * sfeed - RSS/Atom feed reader * irssi - IRC client * gopher - gopher client * lynx - text web browser * feh - image viewer * mpg123 - audio streaming / local mp3 playing * mpv - video streaming * urxvt - terminal * tmux - terminal mutiplexer * vim - text editor * apvlv - pdf viewer * ratpoison - window manager My operating system of choice is now OpenBSD. ## Computing distractions Computers can be a total time-sink and rather than enhancing your life, giving you super-powers to create things and solve problems. They create more problems and frustrations. Don't waste your life trying to find the perfect operating system or hardware setup, you have more chance of finding the holy grail! Be wary of creating too many customizations and be sure that the ones you do make are worthwhile and replicable. Use something like stow to manage configuration files for future speedy deployment. You should aim to be able to install your OS and replicate your setup on any system in less than 30 minutes. This reduces the importance of the hardware. ## Create things If you disallow computing distractions to take hold you should find time to exercise your creativity. Writing texts or maybe software... How you express your creativity is up to you, but don't settle for being a consumer, there is no lasting satisfaction to be had there. Computers should be a means to an end, not an end in themselves. Above all, choose freedom ... [1](gopher://gopher.icu/0/phlog/Computing/Low-power-computing.md) [2](gopher://gopher.icu/I/images/asus3.jpg)

Offline

2023-12-03T12:02:05+01:00

# Offline I have been using a 4G router on the mobile network for the past couple of months and apart from the odd bit of lag it has been quite reliable. This morning however, having unusually left the router switched on all night, my connection appeared to be down. Power-cycling the router seemed to revive the connection, although it wasn't long before I started having issues again. Once more I rebooted the router, it was slow but again the connection returned and I continued with checking my RSS feeds, email and a conversation on IRC, then it went off again ... At this point I gave up and went to get a shower, after that I got my kindle and went down stairs. ## Radio I'm not sure what prompted me to put on the radio, other than it catching my attention from the bookshelf where it occupies a space. I browsed around a little twisting the dial around, first on AM and then on FM, where there seemed to be more choice of channels. While turning the dial I stumbled upon a talk show where a woman was being interviewed about the UK government figures for immigration. She was explaining, in some detail, at how these figures were arrived at and the effects of certain policies on those numbers both historically and going forward. The thing that struck me about this conversation was that this was someone knowledgeable on the situation giving a fair account. This was not news as we are usually fed it today, or the circus that is Westminster, where there is some agenda behind how the information is used to slight some person or party. It was a refreshing experience to hear an expert explain something, seemingly without bias. Maybe I should listen to the radio more often ... This actually reminded me of a video I saw recently about short wave radios, where an American was explaining how he would sit and listen to radio broadcasts from other countries to have a more rounded view of what was actually going on in the world, free from the biases of his own. The BBC used short wave radio for propaganda during the second world war, as they knew the signal could be received deep into Eastern Europe and Russia. It has been reinstated since the beginning of the Ukraine war for this very purpose. You can't firewall radio waves or censor their broadcasts from outside of your jurisdiction. In some countries it was a crime to own a short wave radio for this reason. ## Microadventures I enjoyed reading Jay Scot's phlog post a couple of days ago. It inspired me to get out for a walk yesterday, in the very fresh but unusually dry and sunny weather. Post COVID I had started going out for walks occasionally and had hoped to make a regular habit of it and even maybe expand upon it to do some wild camping. I've spent far too much time watching video howto's not to at least give it a try. Plus I think I already have everything I'd need, other than a bivvy or tarp, to do an overnighter. It felt good to be out yesterday. I definitely need to get out more but I will probably hold off on the overnighter, until the weather warms up a bit! ## Phlogging It's been a while since I posted anything but sometimes I think if you don't really have anything to say then maybe it's best to say nothing... Usually I would write the post on my VPS and just copy it to the relevant directory but that is not possible today, so it's back to old-school offline writing which I will scp across once my internet connection returns.

Phlog-format

2025-02-21T12:46:33+01:00

# Phlog format I almost daily check bongusta[1] to see if there are any new tasty morsels to consume. Generally I also pay attention to the presentation of phlogs as well as their content. Some I have actually stopped reading due to bad decisions regarding line length and being inconsiderate of the wider world of gopher client compatibility. So when someone makes a change that looks good, and causes no issues, maybe there is no shame in copying it? Matto's phlog[2] was the inspiration for the latest format change. You may notice that my article dates are all now nicely right justified. When I saw Matto had done this, it just seemed like the right thing to do. After all, most of us read from left to right and so the most important information should probably come first. Visually it also gives some balance to the page. *Note:* I am still using the UMN gopher client and if your phlog looks like crap (filled with terminal escape codes), or doesn't work (using menus for content to emulate web pages), in that client then you can guarantee I don't read it. [1](gopher://i-logout.cz/1/bongusta) [2](gopher://box.matto.nl/1/)

Usenet-update

2023-12-20T13:10:02+01:00

# Usenet update Some big news regarding Usenet[1]: > Starting on February 22, 2024, you can no longer use Google > Groups (at groups.google.com) to post content to Usenet groups, > subscribe to Usenet groups, or view new Usenet content. You can > continue to view and search for historical Usenet content posted > before February 22, 2024, on Google Groups. In my view googles groups integration with Usenet practically destroyed it. As they do with most things, they assimilate and destroy wholesome technologies from within. ## What's so good about Usenet? There are several good reasons to make efforts to revive Usenet: * It is the original decentralised, federated forum. * It uses a simple protocol designed to work over low bandwidth and impermanent communications links, like dial-up. * News reading clients are ubiquitous and part of many email clients, even if you don't realize it. * The small technical barrier to entry keeps undesirables away, or at least that's how it should play out once google ceases to make it easy for web users to access Usenet. * Forums on websites come and go with the fortunes and interests of their often corporate owners. Because of the federated nature of Usenet, the information therein continues to live on. * ISPs no longer routinely provide access to Usenet. Is Eternal September[2] finally coming to an end? No longer are ISPs routinely providing access to news servers and from February google will stop providing access to people via the web. This should return the sovereignty of Usenet to those who should rightfully have a claim to it. There has never been a better time to reclaim the wonderful resource that is Usenet. ## How do I access Usenet? Please read the wonderful guide[3] written by Matto. [1](https://www.theregister.com/2023/12/18/google_ends_usenet_links/) [2](gopher://gopherpedia.com/0/Eternal%20September) [3](gopher://box.matto.nl:70/0/usenet-getting-started-guide.txt)

Nokia-BLJ-2-battery-refurb

2024-01-13T03:10:48+01:00

# Nokia BLJ-2 Battery Refurb I have a couple of old Nokia banana phones, an 8110 and an 8146 which both have dead batteries. If you should manage to find one of these batteries second hand, they are extortionately expensive. ## Battery Disassembly Dismantling the battery is fairly easy, carefully insert a spudger / guitar plectrum down the seam on the inside of the battery and gently pry it up. There is some black adhesive used to stick the cover to the battery so you may want to get something longer underneath to pry closer to the adhesive until you manage to separate it. Once the outer cover is off you can see the cells and the circuit board. I carefully separated the battery cells from the casing along with the circuit board. To separate the battery from the circuit board I initially gently worked the circuit-board back and forth until the metal battery tags sheared. After this I used a soldering iron to remove the remainder of the pads from the circuit-board and added a little fresh solder for later. ## Parts The BLJ-2 battery is rated at 7.2v and had two cells. I figured if I could find two 3.7v cells of a suitable size then I could wire them in series to achieve close to the voltage I needed. Having measured the available space inside the battery housing, and looking around at various battery packs, I discovered that two 503040 battery packs could be fitted to replace the original cells. Each of the packs is rated at 600mah 3.7v. ## Fitting The battery packs come with a circuit board attached to prevent over voltage, and other things, but this seems to get in the way of using them in series and also there isn't space inside the housing for the batteries with these fitted, so I removed them. I cut the battery tags as close as I could to the circuit board with side cutters, being careful not to short the battery terminals. Once that was done I tinned the battery tags with a little solder and then placed them inside the battery shell along with the circuit board ready to wire everything up. Here I think a picture best explains what has been done[1]. The final assembly is quite snug once the cover is popped back in place. I used no adhesive to secure the new cells or cover and nothing appears loose. [1](gopher://gopher.icu/I/images/blj-2.jpg)

Computing-in-2024

2024-01-19T17:59:52+01:00

# Computing in 2024 Last year I made several changes that reduced the energy consumption of my computing activities at home. This year I plan to provide some of my energy requirements via solar power. I also want to start experimenting with ham radio data modes. ## Computer The Asus B202 eeebox[1] has been serving me well for the past 5 months, though I have had to recently replace the fan and bios battery. It is sufficient for everything but dealing with a full-fat web browser. Ungoogled-chromium works for most sites but not all, with some it really struggles. I've decided that there is already more than enough used I.T. surplus in the world and that something already out there can surely fit my needs. I have recently been thinking of changing the above system for a HP T620[2] thin client. This should provide a small foot print, low energy (15w) 64bit system with no moving parts that would offer a good performance upgrade over the current system. Other than that I'm quite happy with my current setup: * Asus B202 eeebox (1.6Ghz Atom 32bit CPU, 2Gb RAM and 120Gb SSD) * 24" Asus pro-art monitor (1920x1200) * Lenovo KU-1255 USB Trackpoint keyboard ## Software development I will continue to use shell and awk for any small programming projects. Perl is my default language but I've been forcing myself to learn how to use more primitive tools which are to be found on all unix and linux systems. You could argue that Perl is found on all those systems these days, but I think it's bad practice to keep installing programming languages on systems and to make them requirements. What was once all shell scripts is now a mishmash of perl, python, lua and whatever else. I know as well as anyone how strong the desire is to use what you know, but sometimes, for the greater good, it is better to force yourself to convention and broaden your horizons. ## Solar I need to calculate how much solar power I need to produce and store in cells to operate without downtime. This means at some point I will be purchasing solar panels and either an all-in-one solar generator or building my own from components (battery, charge controller, inverter, ...). ## Radio I've been studying for my foundation ham radio license and hope to soon start mixing computing with radio for communications with data modes. I'm looking forward to experimentng with JS8Call, ft8, winlink and RTTY. [1](gopher://gopher.icu/I/images/asus3.jpg) [2](https://support.hp.com/gb-en/document/c04017240)

The-state-of-gopher

2024-02-05T20:31:55+01:00

# The state of gopher I've become a little frustrated with gopher recently. The feelings are mixed because I'm happy there is more content, created by real people. However, many people are arriving in this space and although they bitterly complain about the state of the web, they are bringing ideas and content from the web to gopher. If you want to come to a new space, presumably because you are looking for something different, then please leave the web on the web. It's still there, you can visit any time. Just fire up that fat browser and off you go! Just please don't pollute gopher with it. ## Standards and common practice When you start in radio, or many other disciplines, you are informed of the rules and are expected to listen, to get a feel for the etiquette before you participate. Because it's so easy to fire up a gopher server or get a tilde that provides space these days, it's easy to bypass the learning phase and to jump straight in. The rules, in our case, are laid out in the RFC for gopher[1]. [1](gopher://gopher.32kb.net:70/0/rfc/rfc1436.txt) ### 70 Columns Gopher is from a time where your average terminal had a physical limit of 80 columns. To give some space for the type indicator the RFC states "the display string should be kept under 70 characters in length". Now, modern terminals are capable of more, but by ignoring this standard you make things awful for people, with older systems, or those who view in split screens[2]. It's a slippery slope to becoming like the modern web, where developers don't seem to give a crap about standards, backward compatibility or accessibility. ![2](gopher://gopher.icu/I/images/sog1.png) ### Gopher maps are for navigation, not content There is a habit forming of people misusing maps to create content pages with pseudo in-line-links, like on the web. The problem is that pages become full of info lines, which the original gopher client ignores. So when your content spans multiple pages, the original gopher client cannot page down to see anything after the first page. Another side effect of this is that, if the content is a menu you can no longer press 'd' to download it, as was intended for files. You are breaking functionality of the original gopher client, written by the people who created the gopher RFC. Doesn't that hint to you that maybe it wasn't designed for that purpose and that maybe you shouldn't do it? *clause* - I consider it excusable only for creating applications requiring interaction, as there is no other way to do it, with the proviso that links must exist at regular intervals throughout the page to ensure it remains scrollable in the original gopher client. *Addendum 05/02/2024* The RFC position on non-core item types, which includes 'i', is that the client implementation may choose whether to display them: > 3.5 Building clients > > If a client does not understand what a say, type 'B' item (not a > core item) is, then it may simply ignore the item in the directory > listing; the user never even has to see it. Alternatively, the > item could be displayed as an unknown type. To put your content within a non-core type is therefore not advised. Additionally, there is a commit[3] to the client code which notes that it was a deliberate change to "Skip over type 'i' items". Maybe they foresaw the potential for misuse, or it was already being misused, and decided to address it in this way? [3](https://raw.githubusercontent.com/jgoerzen/gopher/master/doc/client.changes) ### Gopher clients are not terminals Please do not use terminal escape codes to attempt colours and artwork in gopher. They may work in your particular client but generally they don't and look horrendous to the casual user[4] [5]. ![4](gopher://gopher.icu/I/images/sog2.png) ![5](gopher://gopher.icu/I/images/sog3.png) ## Conclusion The saddest part for me is that many of these individuals appear to be in IT or technical disciplines and have some proficiency. Unfortunately they choose either not to read the RFC, blatantly ignore it, or adhere to best practice. Don't be that person ...

Responses-to-The-state-of-gopher

2024-02-08T18:14:32+01:00

# Responses to The state of gopher Thank you all for your kind words and consideration regarding the problems I, and others, encounter using gopher at the present time. Clearly there are a lot of opinions, but the following will always remain true: 1. Vintage hardware has a physical 80 character limit. It is a thing, and a large body of people still use it. Unless you have X running on Linux / UNIX then you are also probably stuck with an 80 character limit. 2. Many gopher clients exist, and are in use, that are written to that RFC specification and will never be retrospectively changed. 3. The UMN gopher client ships with most Linux and BSD systems and may be many peoples first experience of gopher. 4. RFC exist so that we can all play nice together. They were written in gentle times by academics that expected them to be followed. 5. Mutually giving each other permission to disregard RFC, due to age or what has become common and doesn't affect you or your cloister, doesn't make it right. See point 4. If you don't like the limitations of the protocol then please make one or use one that better suits your requirements, Gemini? ---- I would like to take this opportunity to ask the community what clients[1] they are currently using? [1](gopher://gopher.icu/1/poll?poll=1707390454) * All contents 100% gopher friendly

Wordprocessing-with-vim-and-pandoc

2024-05-16T12:35:50+02:00

# Wordprocessing with vim and pandoc It has been quite some time since I have needed to create documents for ordinary folks that expect .docx or .pdf files. Having recently taken on some administration responsibilities for the radio club I am a member of, I find myself again needing to create .pdf documents. ## Abiword At first I installed abiword, along with all its dependencies, which turned out to be quite a lot. I tried it briefly, but after many years using vim to edit text, it felt terribly slow and awkward. ## Markdown Markdown is an excellent minimalist markup that doesn't look out of place in plain text documents. My phlog, as you may be aware, is written in markdown. Markdown was chosen as I knew tools existed to convert it into many other document formats. It can easily be converted into web pages or document types suitable for printing. Pandoc is probably the most widely known of these tools. ## Pandoc On OpenBSD I installed the pandoc package and attempted the conversion of a markdown file: > $ pandoc myfile.md -o myfile.pdf Unfortunately something was missing and the conversion failed with an error. After a short search for the error message, and test installing other packages, I discovered that texlive_base-2023 and texlive_texmf-minimal-2023 were also required. ## Workflow I write the documents in vim, using markdown in the same way that I would for my phlog, and then run pandoc on the file to output it in PDF format. ## Summary I use vim for all my text editing and wanted a solution where I could continue to do that. This solution is fast and avoids any interaction with a word processor, while achieving the same result. I'm quite satisfied with it.

Computing-in-2024-Review

2024-12-31T14:23:34+01:00

# Computing in 2024 Review At the beginning of the year I wrote the original article[1] with some goals for the year. ## Computer Not long after I wrote the article I did in fact treat myself to the proposed HP T620 thin client upgrade. OpenBSD works flawlessly on the hardware and so I have been very happy with it. The system is completely silent, which actually took a little getting used to. ## Software development Some bugs were reported in the quarry[2] gopher search engine and duly fixed. During this process I noticed an issue where stale selectors were not being removed. This too has now been fixed, resulting in faster searches and better search results. I haven't written any new software of any significance this year. For the odd small data processing tasks I have used awk. A few weeks ago I purchased a clean used copy of 'The Awk Programming Language', as I happened to see one advertised for a very reasonable price. I will over time absorb the material and hopefully become a more competent AWK programmer. ## Solar I had planned to build a small solar power system to power my computer and radio gear but I couldn't justify the expense. I ended up instead spending the money on radio equipment. ## Radio I had ambitions to experiment with ham radio data modes. This too went by the wayside not long after I passed my foundation exam in February and started actually talking to people on the air. I have kept my computer and radio air-gapped and rather than exploring data modes via computer integration. I have instead been learning probably the oldest data mode of them all, radio telegraphy. Learning morse code has been a real challenge and after several months of persistent effort, I have finally managed to complete the Koch method of learning morse code. To do this I used a website[3] resource supplemented by an android app called CW Studio. This past two weeks I have started answering CQ calls of stations that are going steadily enough to give me a chance of copying them. It is very challenging and engaging at this point so, I am enjoying the medium as it requires skill on the part of the operator. My equipment is, I think, modest. I have a dual band 2m/70cm hand held transceiver, a dual band 2m/70cm mobile 25w transceiver (which I use as a base station) and a Xeigu G90 20w transceiver for 10m - 160m bands. Two of my antennas I made, one I was given and one I bought. They each have their strengths and weaknesses. Combined they give me a good range of abilities. ## Summary I really didn't want to commit to spending more time staring at a computer screen, typing text to communicate with people and build further dependence on computers. Computers are an impersonal medium. When you can't see or hear the person you are communicating with, part of the communication is lost. Morse code is a skill and a method for data exchange. If you want to really communicate, then talk. This has been my year of radio. [1](gopher://gopher.icu/0/phlog/Computing/Computing-in-2024.md) [2](gopher://gopher.icu/1/quarry) [3](https://stendec.io/morse/koch.html)

Computing-like-1995

2024-12-04T14:42:50+01:00

# Computing like 1995 It's strange sometimes how we go full circle. When I first connected one of my PC's to the internet, back in the mid 90's, I discovered IRC, Usenet, Email and the WWW was in its infancy. Fast forward 30 years, we're almost in 2025, and guess what? I'm using IRC, Usenet, Email and now also Gopher. My interest in computers lead me down a big rabbit hole. Into a career as a perl programmer and web developer and eventually, once out of the rat race, computing returned to being a hobby. Thankfully my career was spent mostly with open source software and operating systems. I got to learn a little bit along the way and not to be easily fooled by the gloss on the surface of software. Once you have seen what lurks beyond some of these shiny web front ends and the infrastructure of some seemingly high status companies, you realize that it is much more about appearance than substance. ## IRC If you want to find groups of people with a similar interest to talk to in real-time then this is the place. There are still numerous public IRC networks out there, and private servers, that you can join and participate in the conversation. ## Usenet The original federated platform. Now that google have relinquished their grasp, and stopped bridging Usenet to the internet via google groups, the Usenet world is a much nicer place. There are literally groups for every possible interest, although many may have not had activity for a very long time. That doesn't mean no one is subscribed to them. I have been surprised a few times when posting to a seemingly dead group. It often appears people are very grateful to have something new to read! Revive this wonderful resource, there is no better time. ## Email What can you say about email? Part of daily life for most of us. I still have a couple of accounts with large providers, as backup and for things I think might generate spam. However, I host my own mail server and have done so for many years. I don't trust third parties to not read my emails. ## Gopher Gopher is perfect for micro-publishing; blogs, news articles or even file distribution. There is narrow scope for defining the appearance of plain text, so appearance is not a great consideration. All my articles are exclusive to gopher, I have no social media accounts, no Gemini capsule and no personal website. ## Commonality What does everything above have in common? They can all be accessed using netcat or telnet. A client is not a requirement to interact with these services, if you learn a little about the protocol. ## The future The future is the past. The command line, a companion that follows my will, with awk as my aide and where all errors are my own. You can visit any time, the port is always open, as long as the lights are on. When will we transcend our teletype emulation and our typewriter keymap? Once upon a time someone had a plan, or 9, but they too became corrupted by the world and their new master. Won't someone pick up the mantle so that we may once again see the light? ``` Technological progress has merely provided us with more efficient means for going backwards. -- Aldous Huxley ```

Display-lines-and-UMN-gopher

2025-02-25T17:30:02+01:00

# Display lines and UMN gopher As Matto described in his recent post[1], the gopher RFC suggests we should keep our display lines below 70 characters. Having read the RFC, and made every attempt to keep my gopher hole within specification, I had arrived at 69 as the magic number for line length. My posts are also wrapped at 69 characters, as it is a comfortable line length to read[2]. I have taken some pride in my gopher hole being RFC compliant, compatible with all gopher clients and having unique content and services. When it was pointed out to me a couple of days ago that my newly added date format was being truncated on an 80 character terminal. I was shocked that I could have made such an oversight and immediately set to put it right. This was at 12am, just as I was preparing to go to bed... Having amended my entire gopher hole to reduce the maximum number of characters to 67, to stop the truncation. It suddenly dawned on me that the report only mentioned the menu items, and if it weren't for the '..' suggesting truncation the entire line would have fitted on screen. I reverted my changes to the ASCII banner and section headings and looked to find the UMN gopher source code[3]. ## Into the rabbit hole It took me some time digging until I found what I was looking for[4]. The *magic* happens inside DisplayTitle(), maxlength is the result of COLS-13. COLS I assume comes from curses knowledge of your terminal capabilities, 13 is what is reserved for the info in the left margin. The margin takes the following form: ' --> [XXXX] ' When accounting for the entire line length UMN gopher reserves 13 characters for the left margin, even though it doesn't always use them. Essentially it calculates whether to truncate your line as follows: line <= COLS-13. If your display line is 69, under 70 as per the RFC, then it calculates a deficit of 2 characters on an 80 character display... I've never had cause to look at the source for UMN gopher before or really critically assess the user interface. ## Initial observations * There is a lot of wasted space within the margin. '--> ' could be replaced with terminal foreground/background invert, to indicate the selected item. This would save 3 or 4 characters in one swoop. * The margin width is variable depending on the index number. I think that this should be fixed as when it expands it pushes the menu title across to the right reducing usable space. ## Summary Is it a bug, is it a feature? Should it it be condemned forevermore, or should it be fixed? Lynx suffers in the same regard when it comes to wasted space and variable width info margin. I actually think it's worse... Even with all its quirks and occasional crashes, it is still my preferred client. ## References [1](gopher://box.matto.nl/0/shorter-lines-in-gophermap.txt) [2](https://baymard.com/blog/line-length-readability) [3](https://github.com/jgoerzen/gopher.git) [4](https://github.com/jgoerzen/gopher/blob/master/gopher/manager.c)

UMN-gopher-revisited

2025-03-02T19:54:10+01:00

# UMN gopher revisited In my previous post[1] I explained the issues that I recently found and what I thought might be the best way to address them. In an effort to resolve the issues, today I wrote a patch[2] for the latest source tree[3]. I don't claim to be a competent C programmer, or have any knowledge of ncurses, but I managed to discover enough in a few hours to make an attempt. The left margin, which contains the menu index, is now fixed at 8 columns. In its current skinny format[4], it provides enough room for indexes up to 999999. There are still problems, as file type suffixes can be appended and I don't think those are accounted properly for truncation either. But at least now that the margin doesn't vary in width, there may be more chance of shaking out the rest of the issues. ## Want to try the patch? * git clone the latest source tree[3]. * cd into the resulting gopher directory * download the patch[2] into the same directory * patch < 0001-Removed-selected-item-indicator-from-left-margin-rep.patch * ./configure * gmake (If you are on one of the BSDs you will need this.) ## References [1](gopher://gopher.icu/0/phlog/Computing/Display-lines-and-UMN-gopher.md) [2](gopher://gopher.icu/0/files/0001-Removed-selected-item-indicator-from-left-margin-rep.patch) [3](https://github.com/jgoerzen/gopher.git) ![4](gopher://gopher.icu/I/files/gopher-menu2.png)

UMN-gopher-revisited-pt2

2025-03-03T22:15:00+01:00

# UMN gopher revisited pt2 ## Once more into the breach There were still some anomalies in text string truncation which I was unhappy to leave, causing me to dive back in to the code for a second time. I found a rather strange routine which would truncate menu titles at the first occurrence of '/'. Rather than truncate from the end of the line it would truncate what it needed after the '/'. It was an unexpected behaviour, so I have removed it. The author should have control over how their menu titles appear. Also lines were being truncated at 'maxlength' and then having '..' appended causing the line to again be too long. I have adjusted the truncation to be 'maxlength-2' to account for this. My original patch has been replaced with a combined one including the above changes. Now that I am a little happier with the client[1], and a patch[2] is available to correct the problems, I have put my menu titles back to 69 characters. Here is a before and after comparison screenshot[3] of current behaviour vs my patch on long menu titles. ## Usenet I started a thread on comp.infosystems.gopher on this topic but unfortunately it soon became more about criticism of the client and its shortfalls (in their opinion). The fact that it ignores pages full of type 'i' is an annoyance to many who like creating pseudo web pages using menus. The fact that it was a deliberate design change by the authors of the reference client and writers of the RFC maybe tells you that they weren't intended to be used that way. ## Summary I've done my bit to iron out some of the bugs. If everyone did this then maybe the world would be in a much better state. Unfortunately people seem largely content with complaining. *Bonus info*: You can change the index display in the left margin by setting an environment variable. The default is ' %6d ', there are 8 columns to fill. $export GOPHERMENUSTYLE=' [%4d] ' $export GOPHERMENUSTYLE=' %5d) ' ## References ![1](gopher://gopher.icu/I/files/gopher-menu3.png) [2](gopher://gopher.icu/0/files/0001-Removed-selected-item-indicator-from-left-margin-rep.patch) ![3](gopher://gopher.icu/I/files/gopher-menu4.png)

DICT-dictionary-server-protocol

2025-03-12T10:49:26+01:00

# DICT dictionary server protocol I have often found myself browsing the web just to look up some word definition. This seems a terrible waste of resources just to find out the meaning of a word, and then I remembered there is a protocol that provides this. That protocol is DICT[1]. As with many protocols from the golden era of computing, it can be utilized without a specialized client. ## Some examples > $telnet dict.org 2628 > Trying 199.48.130.6... > Connected to dict.org. > Escape character is '^]'. Type 'help'. > help > 113 help text follows > DEFINE database word -- look up word in database > MATCH database strategy word -- match word in database using str.. > SHOW DB -- list all accessible databases > SHOW DATABASES -- list all accessible databases > SHOW STRAT -- list available matching strategies > SHOW STRATEGIES -- list available matching strategies > SHOW INFO database -- provide information about the da.. > SHOW SERVER -- provide site-specific information > OPTION MIME -- use MIME headers > CLIENT info -- identify client to server > AUTH user string -- provide authentication information > STATUS -- display timing information > HELP -- display this help information > QUIT -- terminate connection > > The following commands are unofficial server extensions for debu.. > only. You may find them useful if you are using telnet as a cli.. > If you are writing a client, you MUST NOT use these commands, si.. > they won't be supported on any other server! > > D word -- DEFINE * word > D database word -- DEFINE database word > M word -- MATCH * . word > M strategy word -- MATCH * strategy word > M database strategy word -- MATCH database strategy word > S -- STATUS > H -- HELP > Q -- QUIT > . > 250 ok To see a list of all available databases. > show db > 110 166 databases present > gcide "The Collaborative International Dictionary of English v.0.48" > wn "WordNet (r) 3.0 (2006)" > moby-thesaurus "Moby Thesaurus II by Grady Ward, 1.0" > elements "The Elements (07Nov00)" > vera "V.E.R.A. -- Virtual Entity of Relevant Acronyms (February 2016)" > jargon "The Jargon File (version 4.4.7, 29 Dec 2003)" > foldoc "The Free On-line Dictionary of Computing (30 December 2018)" > easton "Easton's 1897 Bible Dictionary" > hitchcock "Hitchcock's Bible Names Dictionary (late 1800's)" > bouvier "Bouvier's Law Dictionary, Revised 6th Ed (1856)" > devil "The Devil's Dictionary (1881-1906)" > world02 "CIA World Factbook 2002" > gaz2k-counties "U.S. Gazetteer Counties (2000)" > gaz2k-places "U.S. Gazetteer Places (2000)" > gaz2k-zips "U.S. Gazetteer Zip Code Tabulation Areas (2000)" > ... To do a basic word search of all databases. > d gopher To do a word search of a specific database. > d foldoc gopher > 150 1 definitions retrieved > 151 "gopher" foldoc "The Free On-line Dictionary of Computing (3.. > gopher > > A {distributed} document retrieval sys.. If you want to script something or just use it from the command line then you can do something like the following: > $echo "d foldoc gopher" | nc -N dict.org 2628 | less ## Conclusion There are many dictionaries available so this is a very useful resource. I thought about writing a gopher interface to DICT, but maybe it's worth learning a bit about the protocol[2]. I may come back to it if there is any interest... ## References [1](gopher://gopherpedia.com:70/0/DICT) [2](gopher://gopher.icu/0/files/rfc/rfc2229.txt)

Stabilizing

2025-06-01T00:35:13+02:00

# Stabilizing Over time my computing setup has changed in many ways. Operating systems, input devices, applications and programming languages. The past 12 months or so it has begun to stabilize. Some of the changes have been purely pragmatic, like when I switched to using the Dvorak key map or stopped using a mouse. Both of these changes were because of a physical need to change due to the onset of RSI. A little research and experimentation lead me to choose a more optimal key map for the English language and to banish the rodent, which was the primary source of my issues. Other changes have been due to changes in philosophy and better understanding what I actually need. Such as using less resource hungry software and so requiring lower power hardware to achieve the same outcome. I have also run into the restrictive nature of libraries and frameworks where you are forced to work in the way established by the author of it and no matter how little of it you need, the rest comes along with it. I tinkered also with OO programming for a while but, once the novelty wore off, it seemed like it actually made more work to adhere to the method than what you gained from it. So that too was abandoned. It's easy to get carried along with a fad when others are excited about it. Quite often these people don't even understand why they are excited about something. They are just carried along by the excitement of others and not wanting to feel they are missing out on being part of something new or something great. Unfortunately the world is full with the voices of salesmen and excitable people rather than those of seasoned professionals. ## Seeking wisdom Some time ago I read The Cathedral and the Bazaar by Eric S. Raymond. Currently I am reading The Art of Unix Programming by the same author. In that book he points to the classic text The Unix Programming Environment written by Brian Kernighan and Rob Pike. I am quite familiar with Brian Kernighan's work having on a number of occasions attempted to work through The C programming language and also having recently purchased a copy of The Awk programming language. I think sometimes if you want to understand something you have to go back to the origin and seek out contemporary sources. ## Battlestation Flat surfaces invite objects. My desk was always a mess with transient things, usually things which required dealing with at some point. In an attempt to prevent accumulation of deferred decisions, I no longer have a computer desk. Using a trackpoint keyboard frees you from constantly moving one hand to and from a mouse and needing a desk to operate it. Dragging a wired or wireless piece of plastic around a desk for the sake of moving a virtual pointer on a screen seems somewhat retarded when you give it some thought. The relatively recent accumulation of radio equipment has required a bit of a reshuffle to accommodate it, resulting in those being stacked on top of each other and a morse paddle cable-tied to the arm of my chair. The HP T620 thin client, which is passively cooled and has no moving parts, stands vertically to the right of the radio's. This is not a curated minimalist workstation. It's an unusual arrangement[1], but it works for me at the present time. [1](gopher://gopher.icu/I/images/battlestation.jpg)

Make-your-own-tools

2025-06-09T10:25:40+02:00

# Make your own tools It's convenient to download some package and all its dependencies for some basic functionality. Rarely do you use or need many of the features of the software and even less of the included libraries and dependencies. Since starting in amateur radio I initially avoided logging software and had been keeping only a paper log. More recently I started manually adding entries to a tab separated file so that I could see if I had spoken to some call sign or other before and on what band etc. There are many packages and websites you can sign up to that provide this functionality, but it's unlikely I would use half of it. I just need to log my calls and a tsv file, which I can search with standard text processing tools, is sufficient. Only you can truly understand what you need from a piece of software and once you understand what you need, often you realize it's actually not that much. ## How it works I have the following in my .kshrc file: ''' # Amateur Radio Log . /home//bin/logbook.sh alias LG='lb_log' alias LF='lb_find' ''' > $ LF DL4ST > 2025-06-07 19:18:11 DL4ST 10.120 30m CW 579 4.. The above shows all contacts for DL4ST. > $ LG G0GSY The above will add the call to the logfile and open it with your $EDITOR falling back to vi if it's not defined. Defaults for band and mode are set using environment variables, LB_BAND and LB_MODE in the logbook.sh[0] file, which also contains the shell functions that make it work. I currently have the variables set to 40m and CW respectively, but you can export them again to dynamically change them. > $ export LB_MODE=SSB When you add a contact the frequency is pre-filled based on the band of operation like so 7.### for 40m allowing you to do a simple search and replace with the exact frequency. It will probably evolve over time, but for now it covers the basics of my needs. [0](gopher://gopher.icu/0/files/logbook.sh)

Phlog-questions-challenge:-technology-edition

2025-04-16T01:08:40+02:00

# Phlog questions challenge: Technology Edition I was tagged by Ben[1] to take the phlog questions challenge. ## When Did You First Get Interested In Technology? Most people of my generation have been raised side by side with a lot of technology, both mechanical and electronic. My father was a mechanic and my grandfather had many interests including electronics. I still remember the drawers full of valves and circuit boards in his shed, which he had scavenged from various unrepairable TV's and radios. My grandfather always had interesting electrical items, colour TV's, video recorders, hifi sets etc. Quite often there were sets in pieces that he was repairing when I visited in my younger years. Unfortunately I was not old enough to understand it at that time, but I think that my grandfather and his various projects were what initially got me interested in technology. 1980's home computers only served to deepen my interest and I still have, and occasionally use, my first computer[2]. This was the beginning of a path which eventually lead me to become a programmer and technology tinkerer. ## What's Your Favourite Piece Of Technology All-Time? Mechanical - The bicycle. I have enjoyed riding, modifying and fixing bicycles since I received a Raleigh Grifter[3] from my my parents one Christmas back in the early 1980's. The bicycle makes it possible for a person to travel a hundred or so kilometres in a day with only themselves as the source of power. Electronic - The radio transceiver. I still find it quite amazing that, without any kind of supporting infrastructure, I can communicate with people in other countries around the world. ## What's Your Favourite Piece Of Technology Right Now? My Xiegu G90 HF radio transceiver[4]. I have been enjoying communicating with people in other countries, learning morse code and also going out on field days to operate portable. ## Name One New Cool Piece Of Technology We'll Have In 25 Years! Thought input devices, keyboards will become a thing of the past. I'm not sure if it should be considered cool or frightening that your brain activity will be interpreted by an electronic device... My hope is that we will have advanced e-ink type displays, which require no backlighting. I can tolerate bad keyboards over bad displays. ## Final Thoughts I have learned over time not to be taken in by all the hype around the latest devices, programming languages and technologies. Mostly they are just repackaged versions of something that already performs the same functional task. I am not enthusiastic about future technology, especially as it becomes more pervasive and intrusive into our lives. Both governments and big tech (is there a difference?) are tightening the screws. The only question seemingly is whether Orwell or Huxley will be proven right. My money is on Huxley... Convenience is the wedge that will separate us from our liberty. I choose technology from a time where its purpose was more benign[5], and where batteries and components could be replaced so that its useful life might be extended. Is there such a thing as wholesome technology? The enjoyment is in the doing, the creating and the solving of the puzzle. If technology puts us in the back seat, there will be no enjoyment or satisfaction in it. We are at every opportunity encouraged to be good consumers... ## References [1](gopher://tilde.pink/1/%7ebencollver) [2](gopher://gopher.icu/0/phlog/Computing/Sinclair-ZX-Spectrum-48k.md) [3](gopher://gopherpedia.com/0/Raleigh%20Grifter) ![4](gopher://gopher.icu/I/images/G90-portable.jpg) [5](gopher://gopher.icu/0/phlog/Computing/Simple-Computing.md)

UMN-gopher-no-more

2025-06-16T01:32:38+02:00

# UMN gopher no more I've been an evangelist for this ancient browser with all its bugs and even, unsuccessfully, attempted to get my fixes added to the main repo on github. I won't go into details but my patch was basically thrown back at me with a bunch of requests for changes I saw no merit in. I refused in no uncertain terms and in response any further communication was blocked by the repo owner. The owner also appears to have no personal interest in fixing any of the reported issues[0]. I've taken some time to think about the situation and the end result is that I have consigned UMN gopher to the great bit-bucket in the sky. There is no future in it. ## Future gophering I had the idea some time ago for a client-less gopher client. Essentially just a filter to de-gopher gopher maps, exposing the links so that they can be accessed via my terminal[1], or display the content of files. I have made a very basic filter[2] which seems to do what I need. Although I will also need to make a script to handle complete gopher URI's and request input for type 7 selectors. Example usage: > $ echo "" | nc gopher.icu 70 | ./degopher.awk | less ## References [0](gopher://gopher.icu/0/phlog/Computing/Display-lines-and-UMN-gopher.md) [1](gopher://gopher.icu/0/phlog/Computing/Smart-terminal-not-client.md) [2](gopher://gopher.icu/0/files/degopher.awk)

UNIX-Primitivism

2025-06-16T16:34:42+02:00

# UNIX Primitivism I have been whittling away at my computing environment for some time. The basic idea being to use only what is required in terms of functionality. This firstly lead to a downscaling from the large resource hungry graphical applications to CLI equivalents. They didn't replace all of the functionality of the originals, but included the functionality I actually used. My window manager too was replaced with ratpoison. I had begun my transition to tiling window management under i3wm and now, committed to that choice and keyboard centric navigation, ratpoison seemed like a natural progression. Once that was done, and all the associated packages and their dependencies were removed, disk space and memory usage were reduced quite considerably allowing me to use lower specification hardware. Then of course I had to consider my own development tool choices. My text editor for a long time has been vim and text editing is a core process for me. Of all the programs I use this is the one I would have most difficulty replacing. Perl had also been my go-to language for most of my personal projects. ## Programming languages Programming languages are like toothpaste, especially scripting languages. Many varieties and flavours exist, but they largely perform the same task. It would be nice to have just one capable scripting language and compiled language. Sadly most systems these days are a melting pot of whatever has been in vogue over the past couple of decades. This makes for an awful lot of dependency baggage. I have abandoned my use of Perl and instead been restricting myself to POSIX shell[0], awk, and trying to write portable code. Which is not as easy as it sounds. Recently I discovered gensub(), which I had used quite extensively, was not part of the POSIX awk specification. ## Base utilities The POSIX standard tool set[1] provides the fundamentals for an extensible and portable computing environment. I think today people are largely oblivious that many of these programs exist and even if they are aware of some. They don't ever bother to learn their capabilities and so never use them. There are entire books written about some of them, like ed[2] and awk[3], that give you a true insight into their capabilities. ## Summary There is a lot of included functionality in any linux/unix base system. Adherence to standards like POSIX and the Single UNIX Specification[4] make for interoperability and portability. The Art of UNIX programming[5] paints a wonderful picture of the early UNIX community, their methods, philosophy and collaborative spirit. I know I am not alone in my quest for a similar environment. ** Make POSIX great again! ** ## References [0](https://www.grymoire.com/Unix/Sh.html) [1](https://en.wikipedia.org/wiki/List_of_POSIX_commands) [2](gopher://gopher.icu/9/files/edtut.pdf) [3](gopher://gopher.petergarner.net/9/Cybertech/The_AWK_Programming_Language.pdf) [4](https://en.wikipedia.org/wiki/Single_UNIX_Specification) [5](gopher://port70.de/9/books/unix/taoup.pdf)

URI-open--plumber-and-xclip

2025-06-18T21:41:18+02:00

# URI open plumber and xclip If you are not blessed with a terminal which can identify URI's and take some programmable action on selection[0], then I offer the following solution: Double clicking on a URI should put it into your xclip buffer. Verify this by selecting some text or URI and then running: > $ xclip -o If the above prints out the selected text or URI then you can proceed to install the plumber[1] and configure it to your requirements. ## Plumber configuration Configuration involves exporting environment variables to determine the application to be used for opening the specific URI/file type: PLUMB_IMAGE= PLUMB_MEDIA= PLUMB_GOPHER= PLUMB_TXTGOPHER= PLUMB_PDF= PLUMB_FILEMANAGER= PLUMB_WEB= PLUMB_TXTWEB= PLUMB_FEED= PLUMB_WAIS= PLUMB_CSO= PLUMB_NEWS= PLUMB_NEX= *Note: * If you are into self flagellation you could instead use xdg-open. Create a key binding to trigger calling the plumber with the contents of the xclip buffer. ## An example for ratpoison Edit .ratpoisonrc and add the following: > bind o exec plumb $(xclip -o) After restarting ratpoison you should be able to Ctrl-t+o to open selected URI's. ## References [0](gopher://gopher.icu/0/phlog/Computing/Smart-terminal-not-client.md) [1](git://r-36.net/plumber)

Gophering-without-a-client

2025-06-21T16:48:55+02:00

# Gophering without a client In a recent article I explained why I'd finally stopped using UMN gopher[0]. In that same article I mentioned my idea for a client-less gopher client and that I'd made a basic filter to degopher[1] menus. This filter exposes the URI's so that I can make use of functionality built into my terminal emulator to call a plumber program which knows how to handle them. *Note:* Configuration of that plumber program, and an alternative solution to both the functionality in urxvt and plumber program, were given in a follow-up article[2]. ## gopen I now have a rudimentary implementation of the final part of the puzzle, the gopen[3] script, which can either be called from the command line or by the plumber program. > gopen gopher.icu If the link contains a search URI then it prompts you for entry. > gopen gopher.icu/7/quarry > Search: For every link you open a new shell is spawned so you back-pedal through your navigation by just exiting less. This may work better inside something like the tabbed suckless application? ## Summary The scripts are very crude and can be improved a lot. The degopher filter I have already discovered doesn't handle 'i' type lines well. It's a very crude way of navigating gopher but it makes all URI's actionable from the command line in a common way and I have taken the same approach with NEX. ## References [0](gopher://gopher.icu/0/phlog/Computing/UMN-gopher-no-more.md) [1](gopher://gopher.icu/0/files/degopher.awk) [2](gopher://gopher.icu/0/phlog/Computing/URI-open--plumber-and-xclip.md) [3](gopher://gopher.icu/0/files/gopen)

urxvt-no-more

2025-06-26T23:46:03+02:00

# urxvt no more I've been a urxvt user for a long time. It supported unicode, I could hide the scrollbars to gain a clean looking terminal, have a blinking block cursor and there were flags to make the launched terminals login terminals so my environment would be set correctly. As I mostly live in the terminal these days, maybe it is only natural that this piece of software comes under scrutiny. urxvt is quite a bit lighter than xterm but usually xterm is already included as part of the xorg package on most systems. You are saving nothing without you manually delete it. Yesterday I decided to take a look at suckless's st (simple terminal). It is minute in comparison to either xterm or urxvt. I had stayed away from suckless software as I don't agree with recompiling for changes of basic settings, like fonts or colours. After looking through some of the patches available however, I found that there was an xresources patch to allow settings for fonts and colours to be taken from the generic .Xresources file. This seemed a very reasonable thing to do and convinced me that maybe it was worth a try. ## Configuration First I amended the config.mk file as I'm running on OpenBSD. You have to uncomment some OpenBSD specific lines in the file and remove '-lrt' from the LIBS line. It wasn't obvious I had to do this, but thankfully the included FAQ covers it. I then downloaded and applied the xresources patch where I ran into more problems. Firstly it failed to compile. I later discovered that the patch applied changes to config.def.h and if you have already compiled the application this file does not get moved to config.h. It doesn't overwrite config.h so that it doesn't stomp on any changes you have made. So while sensible, it's not obvious when I applied the patch why I was seeing errors and it wouldn't compile. Once I had got it to install the colouration was not what I expected. The entire text was the colour of my cursor on urxvt. I later discovered that the numeric colour values used in the patch were not correct looking at those statically set in the header. Once I realized the issue and corrected the values it worked as expected. I made a patch for the patch and emailed the patch author along with an explanation, so hopefully this will be fixed soon. I had intended to add a URI opening patch to match the functionality I had with urxvt but I had an alternative[0] way of opening URI's which was good enough to not warrant the extra patch. ## Summary The st terminal is fast and responsive and my chosen font looks somewhat better than on urxvt. The only thing I feel missing is my blinking cursor. There is a patch for it, but I will give it some time before deciding. I do tend to lose my cursor in large documents and the blinking orange block is good at drawing the eye. ## References [0](gopher://gopher.icu/0/phlog/Computing/URI-open--plumber-and-xclip.md)

OpenBSD-web-browsers

2025-06-27T19:23:47+02:00

# OpenBSD web browsers ## ungoogled-chromium Dependencies: > noto-emoji-20241002 > noto-fonts-24.9.1v0 > nspr-4.36 > nss-3.110 > libudev-openbsd-20230921p0 > libxslt-1.1.43p0 > epoll-shim-0.0.20240608 > wayland-1.23.1 > libxkbcommon-1.8.1 > dconf-0.40.0p2 > adwaita-icon-theme-legacy-46.2p0 > adwaita-icon-theme-47.0p0 > gtk+3-3.24.49 > jtk+3-cups-3.24.49 > xdg-utils-1.2 1.134G of disk usage in my estimation. The noto-fonts are enormous and the major bulk of the installation. I have been using this browser since I started using OpenBSD on my daily driver, around 3 years. There is no Firefox for 32bit, so this was really my only choice and I've stuck with it until today. Performance is acceptable and it seems to cope with most websites. Unfortunately every so often a website will cause it to go into a death spiral which will freeze my system if I don't catch it quickly. ## Vimb Dependencies: > iso-codes-4.17.0:webkitgtk41-2.48.3 > iso-codes-4.17.0:gst-plugins-bad-1.26.0 > iso-codes-4.17.0:gst-plugins-good-1.26.0 > iso-codes-4.17.0:gst-plugins-base-1.26.0 > iso-codes-4.17.0 > mozilla-dicts-en-GB-1.3p1:enchant2-2.8.2 > mozilla-dicts-en-GB-1.3p1:hunspell-1.7.2 > mozilla-dicts-en-GB-1.3p1 > pango-1.56.3:gtk+3-3.24.49 > wayland-1.23.1:libxkbcommon-1.8.1 > wayland-1.23.1 > orc-0.4.32 > lame-3.100p2:libbs2b-3.1.0p5 > lame-3.100p2:twolame-0.4.0 > lame-3.100p2:libsndfile-1.2.2p0 > libunbound-1.22.0:geoclue2-2.7.2 > libunbound-1.22.0:libsoup3-3.6.5 > libunbound-1.22.0:glib2-networking-2.80.1p0 > flac-1.5.0 > libnotify-0.8.6 > hyphen-2.8.8p0 > aspell-0.60.8.1p0 > taglib-1.13.1 > woff2-1.0.2p0 > dbus-daemon-launch-helper-1.16.2 > epoll-shim-0.0.20240608 > libavif-1.1.1 > wavpack-5.6.0p0 > cdparanoia-3.a9.8p5 > json-glib-1.10.6 > soundtouch-2.3.3 > duktape-2.7.0p1:libproxy-0.5.9p2 > duktape-2.7.0p1 > libxslt-1.1.43p0 > harfbuzz-icu-11.0.0 > opencore-amr-0.1.6 > at-spi2-core-2.54.1 > gsettings-desktop-schemas-47.1p0 > gstreamer1-1.26.0 > dbus-1.16.2v0:avahi-glib-0.8p3 > libshout-2.4.5 > libpsl-0.21.1p0 > graphene-1.10.8p1 > flite-2.2 > dconf-0.40.0p2 > adwaita-icon-theme-legacy-4...:adwaita-icon-theme-47.0p0 > adwaita-icon-theme-legacy-46.2p0 400MB of disk usage in my estimation. Dependency listing looks worse than it is with regard to overall size of installation. Unfortunately performance on my system sucked, YMMV. It didn't last more than a couple of minutes before it was uninstalled. ## Firefox-ESR Dependencies: > nspr-4.36 > nss-3.110 > adwaita-icon-theme-legacy-46.2p0 > adwaita-icon-theme-47.0p0 > at-spi2-core-2.54.1 > epoll-shim-0.0.20240608 > wayland-1.23.1 > libxkbcommon-1.8.1 > dconf-0.40.0p2 > gtk+3-3.24.49 300MB of disk usage in my estimation. I have moral issues with the Mozilla Foundation but performance wise, it still seems to be the best of a bad bunch. ## Wayland? Not so long ago wayland wasn't even a thing, now it is seemingly a requirement to significant pieces of software, leverage? Is this what people do when they want to force their beliefs and choices upon others? Redhat with systemd and then Firefox with Rust. Now all web browsers seem to come bundled with wayland, whether you want or not... ## Summary This trip down the rabbit hole was prompted by the OpenBSD 7.7 update. As time goes on it seems to be bloating and the update failed due to my running out of available disk space. This forced me to comb through the system, uninstall anything I wasn't using and check the sizes of the larger packages. That of course meant the web browser. I use vim key-bindings wherever I can get them, so I added the vimium plug-in as I had done with ungoogled-chromium. Emerging from the rabbit hole I'm back on firefox-esr. The performance to bloat ratio is better than the rest.

Legacy

2025-07-10T01:34:42+02:00

# Legacy Software over time never shrinks, it only expands. The legacy tail of hardware on that software is long. We are still emulating teletypes, software support still exists for many terminal types, among other things, that have long since become extinct. We are still using qwerty keymaps, designed to avoid the binding of physical hammers. Horizontally offset keys to allow for the physical bars which connected the keys to the hammers. The list goes on ... The reasons for many of these design choices have long gone, but they are entrenched. Maybe after a time people don't stop to question why? It becomes the standard, the norm and there are too many layers to risk significant change. ## Hardware There is no excuse with regard to hardware. We don't need to use qwerty keymaps or horizontally offset keyboards. The manufacturers just need to change their offerings or we need to find keyboards we can switch the keycaps. Personally I use a remapped qwerty keyboard with a dvorak keymap. I haven't needed to look at the keys for a very long time, even when I was learning I put a printout behind the keyboard so that I could keep my hands in place. ## Software This one is more challenging because of the nature of software. Everything is derived from or built upon something which came before and removing things at the lower levels of the stack can have unforeseen effects on what is above. Plan9 I think tried to sidestep this problem of legacy. Being something new, something different, there were no expectations of backward compatibility, that gives a lot of freedom. Although they did port a lot of their familiar tools from UNIX. ## Cutting the tail I'm not sure if there has ever been an audit of hardware still in use but if not, I think it would be a great help so that there is a point you know you can cut the tail without causing problems. To be able to confidently remove swathes of decades old code would surely be something worth celebrating? ## Old is not necessarily bad With regard to hardware there are tradeoffs. Older is usually less power efficient and performant, but what you sacrifice in power efficiency and speed you usually gain in longevity, repairability and robustness. With regard to software many lessons are learned over time and mature software usually has the majority of the bugs fixed and implementation issues ironed out. Reimplementations can often introduce new bugs and repeat previous, already resolved implementation errors. ## Summary A note to hardware producers, we no longer need horizontally staggered keys, unless the reason is to improve ergonomics. Please also produce keyboards with switchable key caps so they can be arranged for more optimal key maps. Computer users, you should use a keymap that is most optimal for your usage[1]. OpenBSD is meant to be comparatively minimalist and I have very few packages installed, yet I am using 5.7G of my 7.3G root partition. Last sysupgrade I ran out of available space to complete and the situation, as explained above, will only continue to get worse. How do you identify and remove legacy code from the current stack without it all coming crashing down? ## References [1](http://patorjk.com/keyboard-layout-analyzer/v2/)

ed-vi-vim-sam-vis-editor

2025-07-12T01:39:37+02:00

# ed vi vim sam vis editor I have been on a bit of an editor tour over the past week or so, after being initially inspired to look more closely at ed. Getting to know ed a little has been a bit of an enlightenment to me. As a long time vim user I was left feeling that I had been using it all wrong! Going further down the rabbit hole lead me to install and do some experimentation with sam[1]. I hated the GUI, but with the -d flag you can use it much like ed on steroids. Here is a quick reference to its command language[2]. ## Structured regular expressions Once you have seen these[3] in action you can not simply continue to use an editor without. Rather than working with lines you are working with precicely constrained regions of your document. Only a demonstration[4] can do it justice. This eventually lead me to looking for some happy middle ground between the minimalism of ed, the functionality I had become accustomed to with vim along with the power of sam. ## vis vis[5] is not a fork of vi or vim. It is a ground up reimagining of vi combined with the power of sam's command language. It feels more like a UNIX tool, even the text reformatting is a call to '|fmt'. Something that always bothered me about vim was that, modal editing could be enjoyed while working on your document, but in the command mode you were back to plain old typewriter. This is not the case with vis, which allows you to modal edit your command line. This is a real boon for me as I use vim bindings and modal text entry wherever I can get it. There are some things which have bothered me, but it's too early to tell if these are long term issues or just teething problems because it is not vim. It is however a fraction of the size of vim and proving to be a very capable editor. ## Historical knowledge For those who were there at the beginning and learned ed while using teletypes a certain way of working became established and when later tools came along those established ways developed along with the tools. For people that have come along since, maybe coming from a different paradigm for editing text such as using WYSIWG word processors and editors, we come with baggage. I learned to use vi, an then vim, by myself without anyone to give me pointers and that has resulted in my use of it being very much as a visual editor with lots of hjkl scrolling around to make changes to parts of the text as you would in a word processor. ## Conclusion Had I started with ed the story would have been very much different. Having spent just a little time learning ed, it has made me aware of the following. I very rarely made use of the search feature to jump to words, line ranges to constrain or widen the scope of my commands , regex search and replace or the ability to send parts, or all, of a file for processing by external programs. And this is where the real power lies. PS. This document was produced in vis. ## References [1](http://doc.cat-v.org/plan_9/4th_edition/papers/sam/) [2](http://sam.cat-v.org/cheatsheet/) [3](http://doc.cat-v.org/bell_labs/sam_lang_tutorial/sam_tut.pdf) [4](https://www.youtube.com/embed/VZo_F60oqbI) [5](https://sr.ht/~martanne/vis/)

password-store

2025-07-18T00:38:20+02:00

# password store A while ago I read a post by Jay Scot[1] that described the numerous dependencies of the pass(1) password-store on an OpenBSD system and his replacement of it with a shell script and a command to generate new password files. His script is tailored to work with dmenu. I have largely plagiarized his work and modified it for my own purposes. In doing so I have integrated the new password generation command into the shell script[2]. ## Usage You must first export an environment variable for your password store. If you have been using pass already on OpenBSD, then the following should work: > $ export PASSWORD_STORE="$HOME/.password-store" Don't forget to add the above export to your .kshrc, or whatever, so that it is set next time you come to use it. If all is well the following will list all your password files. > $ pw The following will list files based on a partial match, if there are more than one, or it will request you to enter your password to decrypt the file. > $ pw If only one was found, after entering the password, it will be copied to your pastebuffer for pasting. ## Flags There are two additional flags that can be used: -n Which when supplied with a filename will generate a random password, create a new password file and copy the password to your pastebuffer. > $ pw -n -v Which will list files based on a partial match, if there are more than one, or it will request you to enter your password to decrypt the file. If only one was found, after entering the password, the file will be printed to stdout. > $ pw -v This is useful if you have several pieces of information within the file you need access to. At the top of the pw script is a variable, PB_SECONDS. This determines the amount of seconds after which the pastebuffer is overwritten, the default is 30 seconds. ## References [1]( gopher://ams.jay.scot/0/phlog/016.txt ) [2]( gopher://gopher.icu/0/files/pw )

rc-shell-and-new-gopher-server

2025-07-26T02:03:48+02:00

# rc shell and new gopher server I had been dipping in and out of the Plan9 rabbit hole, without going too deep. If you had been following along recently, then you know I had also been experimenting with ed and sam which lead me to install vis, which shares sam's command language. In the process my OpenBSD system has become a bit of a Franken-9, with rc as the shell on my local system and installed on my server. ## rc shell After using perl for many years I find POSIX shell quite limiting, especially the lack of lists/arrays. So when I discovered the rc shell has lists, much cleaner syntax and had dispatched a number of bourne keywords determined to be redundant. I decided to give it a try. There were three options in the OpenBSD repos, ordinary, and with editline or readline: > 1: rc-1.7.4p1 > 2: rc-1.7.4p1-editline > 3: rc-1.7.4p1-readline As I like modal commandline editing I gave both of the later a try. I settled on the readline variety as it seemed more compatible with what I had become used to with ksh's vi-mode. I started by creating an .rcrc file and converting my existing .kshrc to rc. This was a very helpful learning experience to start to familiarize myself with the syntax, as was reading Tom Duff's paper on the rc shell[1]. I modified my prompt to include the directory path using the examples from the paper. Then in short order I converted my aliases to functions and ported the remainder of my ksh functionality. By comparison the resulting file looks a lot neater. ## gopher server (gophrc) Encouraged by what I had learned so far, I wanted to continue familiarizing myself with the language. The best way for me to do that is be creating some small project. I am of the type that learns through doing and a small project is usually how I go about learning. I remembered I had started writing a gopher server in shell that ran from inetd but had run into some issues and mostly given up on the idea after discovering Katolaz had already created one called gosher[2]. So, for the past couple of days I have been chipping away at making *gophrc*[3] a gopher server that runs from inetd and that can work with my existing geomyidae formated menus and cgi programs. You are reading this file from that experimental server. ## Summary It has been a great learning experience and I am a great fan of home brew software. I feel that rc shell has a lot of potential, but that I am still missing a lot regarding nuances of the language. If you know of any more documentation than that I've linked below, then please drop me a line. If you spot anything that doesn't work or things I could have done better in rc shell, then likewise please let me know. ## References [1]( http://9p.io/sys/doc/rc.html ) [2]( gopher://katolaz.net/1/software/gosher ) [3]( gopher://gopher.icu/0/files/gophrc )

The-UNIX-way--a-transcript

2025-08-11T22:01:21+02:00

# The UNIX way a transcript To illustrate the UNIX method of using small programs dedicated to specific tasks and where output of one is the input to another. The following example shows creating and editing a dynamic playlist using four distinct programs: * ls (list files) * rlwrap (provide searchable command history and command line editing) * ed (the standard UNIX text editor) * mpg123 (mpeg music player) Here is the transcript: hotaru$ rlwrap ed '!ls -1 /home/archive/media/Music/Soul\ II\ Soul ... 1559 ,n 1 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 2 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 3 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 4 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 5 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 6 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 7 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 8 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 9 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 10 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 11 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 12 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 13 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 14 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 15 /home/archive/media/Music/Soul II Soul/Volume IV_ The Clas ... 13,15d w !mpg123 -@ - High Performance MPEG 1.0/2.0/2.5 Audio Player for Layers 1, 2 and ... version 1.32.10; written and copyright by Michael Hipp and ... free software (LGPL) without any warranty but with best wi ... Directory: /home/archive/media/Music/Soul II Soul/Volume IV_ The C ... Terminal control enabled, press 'h' for listing of keys and functi ... Playing MPEG stream 1 of 12: 01 Back To Life (However Do You Want ... MPEG 1.0 L III cbr320 44100 stereo Title: Back To Life (However Do You Want Me) ... Comment: 1 ... Year: 1993 ... ## Summary Only once you start thinking in this way can the true intention and power of UNIX be realized.

Gopher-cleaning

2025-07-27T02:05:15+02:00

# Gopher cleaning It has been some time since I gave my burrow a good cleaning. By cleaning, I mean looking through a lot of the programs which drive the phlog and other gopher services hosted here. Changing the server software gave me reason to start poking around and, as often happens. You start finding bugs or other things that in your present mind you can not accept. This leads to fixing things in such a way that the present you can reconsile. Before you know it you are deep in the rabbit hole. There was an issue with the new gopher server logging, it was mangling the ipv6 addresses due to my poor initial attempt to split on ':'. The polling program too had some problems, it should have been ordering the polls by most recently active, but it wasn't. There really is too much to remember and list but I should probably leave things to settle for a few days before I go digging in there again, as maybe I have broken other things in the process. I have also been at it for the majority of the day... If you find anything broken then please let me know.

Timezones

2024-12-29T01:07:13+01:00

# Timezones and daylight saving Do we really need them and do we really need ones which change? Logically thinking, if there was a world time which didn't change, then appointments, schedules and whatever else should always be possible to synchronize internationally without any ambiguity. The military uses UTC (GMT or Zulu) time which remains constant, radio amateurs also use this singular zone when logging or arranging schedules. ## Why am I even talking about this? My server is located in Amsterdam and the VM is defaulted to CET, my local machine is GMT. Both of these change dependant on time of year. GMT switches to BST (UTC +1hr) at 01:00 on the last Sunday of March until 02:00 the last Sunday of October were it returns to GMT. I'm fairly sure CET follows the same schedule but is GMT +1 to begin... I send my email from the server so it has the server time. I also sync files between the server and my local machine so it's obviously much better if their clocks are in sync. This led me to the following questions: * Why do they change? * Do they need to change? * What are the advantages and disadvantages? * Is there a better way? ## Would it be weird your morning starting at 13:00? I guess we have gotten used to the small hours being the very early morning and the larger hours being the evening. But would it really make any difference? Would the benefits of a common global time outweigh the perceived problems? How much computation is performed just converting dates from one zone to another? ## What really matters? I think what really matters are daylight hours. There have been various studies[0] performed by the British government over time and statistical analysis by safety bodies like RoSPA that strongly suggest that with extended daylight hours there are less road traffic deaths. Another benefit is that less electricity would be consumed keeping the lights on during darkness. One campaign, Lighter Later, suggested a change to increased daylight hours, GMT+1 in winter and GMT+2 in summer, could "save almost 500,000 tonnes of CO2 each year, equivalent to taking 185,000 cars off the road permanently". That's quite a claim. Why can't we just change the alarm clock and not the time? [0](gopher://gopherpedia.com/0/British%20Summer%20Time)

Down-the-rabbit-hole--9front

2025-08-18T00:23:56+02:00

# Down the rabbit hole 9front As time has passed and frustrations with the enormity and complexity of software have grown. So too has the motivation to look for alternatives. Recently I purchased a tiny Dell 3040 in the hopes of running either Plan9 or 9Legacy. Unfortunately the dell appears to be UEFI boot only. There is mention of a legacy boot option, but I am unable to find it even after downgrading the firmware to the earliest version I can find, 1.2.0. I did get 9front to boot, but could not install as the onboard MMC didn't appear as a usable device. Not wanting to give up, I first tried 9Legacy on the Acer B202 I still had laying around but it failed early in the boot process. I then tried 9front, which I successfully managed to install with both networking and sound devices appearing to be recognized. ## Installation The usb stick install media is like a live CD. Once booted and inside rio, if you need to change the keymap then run: > % /bin/kbmap Right click to select the keymap, then press 'q' to quit. To begin installation: > % inst/start configfs - will ask you to choose a filesystem type, for my purposes it was recommended gefs (good enough file system). partdisk - you can delete and create a partition for your system. Here I ran into a problem with the following stage 'prepdisk' which is meant to divide the allocated partition. It failed repeatedly until I left a couple of sectors at the beginning of the disk free during this stage. So, I couldn't allocate from 0 - . prepdisk - automatically divides the partition or you can manually override. I just went with what it recommended. mountfs - reams/formats the partitions and mounts them configdist - choose the location of the install media confignet - choose manual/auto (dhcp). tzsetup - choose timezone (GMT) bootsetup - install mbr and mark partition active (yes to both). finish - reboot ## Post install First thing I needed to do was make my keymap selection permanent. To do this edit $home/lib/profile, depending on your level of bravery your choices of editor are ed, sam or acme. Be ware, there are different sections in profile which are run depending if you are logging in from a local terminal or as a cpu. I made the mistake of adding the line to set the keyboard at the very top of the file and it really didn't like it when I connected to it using drawterm from another computer. Make sure local changes go under the terminal section. To get a list of keymaps: > % ls /sys/lib/kbmap Add the following line to your profile under the terminal section: cp /sys/lib/kbmap/ /dev/kbmap ## ssh to other systems > % auth/rsagen -t 'service=ssh' >$home/lib/sshkey > % auth/rsa2ssh $home/lib/sshkey >$home/lib/sshkey.pub > % cat $home/lib/sshkey >/mnt/factotum/ctl I then copied the public key to .ssh/authorized_keys on my remote machine. *WARNING* Do not try to ssh to a remote machine from an ordinary 9front terminal! It hung and crashed rio in my case. For vt emulation there is a vt program which you must start before trying to use ssh: > % vt -xb Then: > % ssh user@remotehost I will probably create a function in my profile to make this override ssh with a single command. ## Summary This is as much a reference for myself as it is maybe of some use to someone else. My memory isn't great so having something to refer to, should I have to do it again, is quite useful. The current state of play is that I have a 9front system that I can connect to using drawterm from my usual openbsd desktop and I can connect from the 9front system to my vps. Old habits die hard. It is evident, when it's not there, that I use less a lot. What is significant about the 9systems is not what is included, but what has been removed. less/more is a good example of this, you don't need a pager if your terminal window behaves as a pager. Terminal windows by default do not scroll to the bottom like a unix terminal would, so man pages etc appear to just be cat'd onto the screen. This small adjustment from using less to using cat and the screen not scrolling is taking some getting used to, and this is just one small illustration. Coming from linux/unix it is frustrating because while you know it is different, it looks very familiar and even some of the same programs exist, yet don't quite behave the same. If you do get stuck without less and it becomes too much there is 'p' which is a pager. It is tempting to alias it, and a number of other things, to more familiar names. But then you would be convincing yourself more that it is what it is not. Much of the above install and configuration information was gathered from the 9front Wiki[1] and this excellent Plan 9 Desktop Guide[2]. ## References [1]( https://wiki.9front.org ) [2]( https://pspodcasting.net/dan/blog/2019/plan9_desktop.html )

Down-the-rabbit-hole--9front-pt2

2025-08-20T03:18:15+02:00

# Down the rabbit hole 9front pt2 It's a heavy time investment learning a new system and unlearning old habits. Today however I feel I have made a little progress. ## No usable web browser I've overcome this problem by installing a light linux distribution to a tiny dell 3040 thin client to which I've installed a vnc server and firefox-esr. This works a lot better than I thought and does away with the need for the usual solutions of running 9front from within a vm or using drawterm from some other OS to operate the networked 9front system. I wanted to be on bare metal using the system natively so I wasn't prepared to accept those crutches. ## Command history One of my early complaints was lack of command history. I was fairly sure that I must be missing something as this would be quite an oversight. To recall the last command: > % " To run the last command again: > % "" Now for the interesting part. To recall a command matching a pattern: > % " To run a command matching a pattern: > % "" If you're not satisfied with that, then how about this. All the text you have in your window is in /dev/text. Want to list previous commands? > % grep term% /dev/text *NOTE* Each window is its own entity with its own environment and /dev/text so you will only have contextual history in that window. If you mount some resource in another window, don't expect it to be available in another. ## Keyboard remapping I use modal editing not only in my editor but also on the commandline. Due to this I use the escape key quite heavily and have in recent years got into the habit of remapping caps-lock to escape. Even though I don't use modal editing in 9front I still use it on my server and so want the additional escape key where I'm used to it. In order to do this I had to first discover what keys were being pressed. 9front does not seem to use familiar key codes in its kbmap files. In order to discover the codes in the correct format I had to use evdump(1) which appears to be very similar to xev under xf86. Once I had learned that escape was '0 1 0x1b' and caps-lock was '0 58 0xf017'. I looked to figure out how to add a mapping for capslock to escape. Looking inside /sys/lib/ascii told me what I needed to know '0 1 ^['. So all that was required was to add '0 58 ^[' to my modified kbmap. I had copied the dvorak map to dvorak-riow, so as not to destroy the original with my tinkering. To check if the new file worked, all that was required was to: > % cat /sys/lib/kbmap/dvorak-riow > /dev/kbmap Sure enough caps-lock now behaved as escape. With that confirmed I edited my user profile to reflect the change of keymap for next login. ## Keyboard shortcuts ctrl+a - jump to start of line ctrl+b - jump to text output point (useful to enter next command) ctrl+e - jump to end of line ctrl+d - seems same as linux/unix and closes terminal ctrl+f - autocomplete file names ctrl+h - delete letter backwards ctrl+w - delete word backwards ctrl+u - delete text from cursor back to the output point ## Still To-Do * Map a key to Kmod4 to use riow (I have no windows key!) * Check sound, fix if not working. (I can see audio device) * Change rio window colours (I like white on black) * Get full resolution (1920x1200x32), currently at 1600x1200x32 vesa. * Start winwatch and riow on login winwatch is a taskbar where you can minimize your windows.

Down-the-rabbit-hole--9front-pt3

2025-08-23T21:53:03+02:00

# Down the rabbit hole 9front pt3 I have worked a little through my To-Do list from last time. ## Keyboard remapping for missing windows key I needed to reassign a key for Kmod4 due to having no windows key. Initially I was going to use Alt but was told it was used as the compose key for typing special characters. Firstly I remapped Alt-gr. But when I actually came to use it, there were functions where you had to hold the button and use the arrow keys, which was practically impossible. In the end I settled upon remapping Alt to Kmod4 (windows key) and Alt-gr to Alt. This feels a lot more natural as the buttons are practically where they were intended to be if I'd had a windows key: > 0 56 0xf868 # alt to windows key > 2 56 0xf015 # alt-gr to alt ## Getting sound working Fairly easy fix from reading the manual. The headphone socket was easily identifiable, once that had been enabled I had sound and near deafened myself. Echoing a more reasonable value into /dev/volume reduced the sound to a more sensible level. Both settings were added to my $home/lib/profile: > echo pin 27 > /dev/audioctl > echo 70 > /dev/volume ## Change rio windows to white on black Today I was reading through a 9front added feature list and discovered that rio had a flag to do this. It was simply a matter of adding -b to the rio startup line in my profile. ## riow and winwatch on login Now that I once again have virtual desktops I don't feel I need winwatch. winwatch was really just a crutch to save me being overwhelmed by too many windows and losing everything. The past few days have been a good reminder why I prefer tiling window managers and virtual desktops. They really do make keeping your work space organized much easier. ## Screen resolution I did a little digging today and I think my graphics card is supported by igfx. The output[1] seems to suggest that igfx should allow me full resolution at 1920x1200. However, on trying to initialize it, I get this error: > term% aux/vga -m igfx -l '1920x1200x32' > aux/vga: main: igfx@1920x1200x32 not in /lib/vgadb ## Summary There is a bit of satisfaction, and relief, to actually now use the system for doing something useful. This text was written in acme to a 9fs filesystem, mounted over ssh, from my OpenBSD server. No longer do I need to be connected to the server via ssh to edit the files directly or scp the files across from my local system. The process now is as simple as writing a file to a local filesystem. The mail attachment downloads directory is also shared in the same way, so that I have easy local access. No more scp downloading email attachments! [1]( gopher://gopher.icu/0/files/output-of-aux-vga--vga.txt )

Down-the-rabbit-hole--9front-pt4

2025-08-29T11:46:59+02:00

# Down the rabbit hole 9front pt4 The To-Do list from my second post in this series has been completed, other than managing to get my monitor to run at its native resolution. Unfortunately now that I've been using 9front for a few days there are some things I have as yet been unable to do and so the To-Do list expanded to include the following. ## Mounting external disks I have a 500Gb archive drive which holds my music and other files. The file-system is ext2 and I wanted to access my music. I also wanted to be able to use usb sticks. As it turns out dos/fat32 formatted USB sticks will auto mount under /shr For ext4 there is ext4srv(1). Firstly you have to start the ext4srv service. To identify your drive and partition, run the following before and after connecting it: > % ls /dev/sd* Then simply: > % mount /srv/ext4 /n/ /dev/sd/linux I added the following to my lib/profile: > ## Archive HDD > ARCH_DRIVE=/dev/sdU1c3b2/linux > if (test -e $ARCH_DRIVE) { > ext4srv > mount /srv/ext4 /n/archive $ARCH_DRIVE > } ## Ham radio logging script migration Even though I had ported my ksh script to rc in advance, it seems that I was still using linuxisms: * No printf on 9front, used echo instead as no formatting needed. * date formatting is quite different: > date -uf 'Y-MM-DD hh:mm:ss' vs > date -u '+%Y-%m-%d %H:%M:%S' I think you'll agree that the date formatting in 9front is cleaner. ## password-store This was mild torture. Due to already having migrated the ksh script to rc on my OpenBSD system, I was expecting a fairly painless migration, but no. Things that were valid syntax on the OpenBSD implementation of rc were not valid on 9front. I had lots of errors about concatenation of null lists, or something like... I found that using $"var instead of $var seemed to alleviate the errors and get things working. I also had to replace sending data to xclip with /mnt/wsys/snarf and use an alternative to gpg to encrypt and decrypt[1] the file. ## Summary The only way to truly learn another system, and discover if it can meet your needs, is to use it how you intend to use it and do the things you would ordinarily do day to day. ## References [1]( https://pspodcasting.net/dan/blog/2019/plan9_desktop.html#pim_pass )

Confession

2025-10-09T12:08:51+02:00

# Confession My adventure down the rabbit hole was short lived. Using a secondary device so that I could access a web browser that would work on modern websites became a significant annoyance, as did rc shell's handling of quoting and non-completion of program names. I also hate the terminal windows and often find my cursor some place other than where I want it, yes I learned about ctrl-b? to get me to that point and a host of other incantations. But collectively it was all just too much when I had a system which was complete and tolerable to hand. So I'm back on OpenBSD, using rc wrapped in readline as my shell, except when I find it too annoying. I also finding myself using vim again rather than vis. I am writing this using vis but I find vim's spell checking integration nicer than piping the document through spell which just presents me with a list of misspelled words that I have to hunt down and modify. I'm sure I could quite easily script something to highlight the offending words. Maybe I should do that and would feel a little happier. riow's key bindings probably made the experience last longer than it would have otherwise. I find the constant use of the mouse tedious and unnecessary. I also dislike the inconsistency between acme and rio, acme has its own way of doing things which in some cases is quite at odds with rio. In the end I preferred to use sam rather than acme as an editor partly due to this. So there we have it, my confession and my reasons. ## Summary It's a real pity as the underlying principles of the operating system are sound and it does fix and simplify a number of UNIX short falls. But like many things that are sound in principle, in use they can have problems. None of them individually unsurmountable, but there are just too many annoyances for me.

Electrolytic-Capacitors

2025-11-13T01:09:19+01:00

# Electrolytic Capacitors Usually the first thing to fail on your electronics devices, after the batteries, are the electrolytic capacitors. Over time they dry out and go out of specification, which can cause damage in itself. However, the tops can mushroom and even split, resulting in acidic goop leaking all over your precious circuits. If left, it can cause a lot of damage to the PCB and other components. Most often when something has stopped working I will open it up and visually inspect the capacitors. Usually I can find two or three obvious candidates. Even if you can't identify the problem ones visually, sometimes they give off a tell tale high pitched whine. Electrolytic capacitors are very cheap and easily obtainable. I've repaired a lot of devices just by replacing a few capacitors, costing only a few pounds. The reward is the satisfaction of reviving your device, or someone else's, giving it a second life and keeping it out of landfill. Today I revived my old Dell Optiplex GX620. I'd taken it apart and bought the replacement capacitors months ago, but today I finally got around to replacing them. The GX620 is now working again, cleaned and reassembled ready for a new owner.