00:00:00 --- log: started forth/19.06.28 00:04:50 --- join: bb010g (bb010gmatr@gateway/shell/matrix.org/x-gktbppjexuaxcyvr) joined #forth 00:13:39 --- join: dys (~dys@2003:5b:203b:102:226:5eff:fee9:68d2) joined #forth 00:51:00 --- join: jimt[m] (jimtmatrix@gateway/shell/matrix.org/x-rllzxhzheiniqpwo) joined #forth 00:51:00 --- join: siraben (sirabenmat@gateway/shell/matrix.org/x-xtshdkqllgtqpzqr) joined #forth 01:15:09 --- join: xek (~xek@37.248.253.114) joined #forth 01:29:42 --- quit: deesix (Ping timeout: 252 seconds) 01:30:14 --- quit: dddddd (Ping timeout: 268 seconds) 01:32:26 --- join: deesix (~dddddd@unaffiliated/dddddd) joined #forth 01:39:03 --- join: dddddd (~dddddd@unaffiliated/dddddd) joined #forth 02:20:37 Heh. Calibrating the temperature sensors has spilled out of production and testing and is starting to absorb R&D people. :D 02:21:08 This one product has so dominated the company it's comical. :D 02:46:13 --- quit: dddddd (Remote host closed the connection) 02:49:28 the mcu inbuilt temp sensor ? 02:59:41 --- quit: xek (Ping timeout: 258 seconds) 04:08:13 they don't have some calibration test sets for doing this semi-automated? 04:11:08 I'm guessing it's probably their IR temperature sensors 04:11:46 no one gets excited about the lame and inaccurate temperature sensor in a mcu 04:19:49 --- join: reepca (~user@208.89.170.37) joined #forth 04:39:07 the built in temp sensor in a mcu is for "its too damn fucking hot or cold" kind of scenarious 04:39:39 ttmrichter: they are feeling the heat but want to know it preciesly? ;3 04:41:05 Zarutian, yes, true re the mcu internal temp sensor 04:41:54 Zarutian, ttmrichter has developed a very sophisticated thermal camera for powerstation work 04:46:51 --- quit: kori (Quit: WeeChat 2.5) 04:47:38 --- join: kori (~kori@2804:14c:85a3:81b8::1001) joined #forth 04:47:49 --- quit: kori (Changing host) 04:47:49 --- join: kori (~kori@unaffiliated/kori) joined #forth 04:57:32 I didn't develop it. 04:57:47 I did the laser rangefinder component of it and the communications switchpoint aspect of it. 04:57:49 That's it. 04:59:00 well youre part of the team that developed it 04:59:34 and the low power aspects ? 04:59:52 As for the complexity: calibrating it involves pointing the thermal camera at a 3x4 array of temperature targets, reading off those points, and then correcting for the real temperature. 04:59:57 No low power aspects on that one. 05:00:19 black body calibration ? 05:00:37 ttmrichter: always been curious. How does an laser rangefinder work, roughly? 05:01:09 ttmrichter: is it based on time of flight or trigonometry like sextants use? 05:01:20 Zarutian: There are several approaches, but the most common is a simple echo. You have the laser interrupted with a coded pattern and time how long it takes for that pattern to come back. 05:01:54 More sophisticated ones will use that as a baseline and then let you figure out heights and such using trigonometry, though, yes. 05:02:03 I have a 10-buck handheld one that does that. 05:02:11 Out to 40m. 05:02:32 amazing how cheap they have become 05:02:43 You point the laser at the top, trigger, point at level (it has a built-in bubble for that), trigger, point at the bottom, trigger. 05:02:57 * Zarutian has seen an exercise where an sextant was used to figure out how distant a containership was from the viewer. The measuring target? One of the containers on the ship. 05:02:57 And from that it calculates the three ranges, does simple trig work, and tells you the height. 05:03:28 Yeah, similar technique, though faster and easier. :D 05:03:45 If I wanted to pay 60 bucks I could do the same thing out to about 200m. 05:04:10 But ... uh ... no. I don't want a laser with that kind of horsepower in my hands. 05:04:30 the idea with the exercise was to teach the use of sextant and application of trigonometry 05:04:41 Yep. 05:04:48 It's a good exercise too. 05:07:36 heard there was an sextant-ish simulator smartphone application one could use to measure angles that fit into the field of view of the camera of the phone 05:11:29 and there is, qlone, an photogrammetic 3d scamming app that I suspect uses similiar techniques. 05:12:19 though one needs an eye-que card or other patterned flat surface underneath the object being scanned for reference 05:54:54 I need to buy myself a real sextant. 06:40:47 --- join: dddddd (~dddddd@unaffiliated/dddddd) joined #forth 06:56:18 --- quit: dave0 (Quit: dave's not here) 07:48:34 --- join: xek (~xek@37.109.33.137) joined #forth 07:49:06 --- join: proteus-guy (~proteusgu@2403:6200:89a6:8231:6809:905b:b461:e8b9) joined #forth 07:49:33 --- quit: proteus-guy (Remote host closed the connection) 08:01:02 --- quit: tabemann (Ping timeout: 276 seconds) 08:05:50 --- quit: dys (Ping timeout: 264 seconds) 08:27:04 --- quit: xek (Ping timeout: 245 seconds) 09:13:27 --- join: rprimus (~micro@unaffiliated/micro) joined #forth 10:10:44 --- join: cnidario (~aaa@92.57.58.87) joined #forth 10:11:21 --- join: mark4 (~mark4@208.54.86.161) joined #forth 10:25:35 --- quit: cnidario (Remote host closed the connection) 11:18:22 --- quit: mark4 (Read error: Connection reset by peer) 11:18:41 --- join: mark4 (~mark4@208.54.86.161) joined #forth 11:20:55 --- join: mark4_ (~mark4@208.54.86.161) joined #forth 11:21:05 --- quit: mark4 (Read error: Connection reset by peer) 12:22:56 --- quit: gravicappa (Ping timeout: 246 seconds) 13:29:02 --- quit: lchvdlch (Read error: Connection reset by peer) 13:34:13 --- join: lchvdlch (~nestr0@191.98.151.137) joined #forth 15:00:50 --- nick: mark4_ -> mafk4 15:00:53 --- nick: mafk4 -> mark4 15:24:50 <`preside1> morning forthnighter 15:41:40 --- join: dys (~dys@tmo-123-188.customers.d1-online.com) joined #forth 16:09:19 --- nick: `preside1 -> `presiden 16:23:48 --- join: dave0 (~dave0@069.d.003.ncl.iprimus.net.au) joined #forth 16:24:34 hi 16:25:28 --- quit: john_cephalopoda (Ping timeout: 252 seconds) 16:37:21 --- quit: mark4 (Ping timeout: 268 seconds) 16:38:49 --- join: john_cephalopoda (~john@unaffiliated/john-cephalopoda/x-6407167) joined #forth 16:44:32 Zarutian: Forþ, a 4 letter word. 16:45:01 pointfree: wrong letter. It should be Forð 16:45:26 þorn is never at the end of a word or word-stem 16:46:30 and Forð is the first four letters of Forði (which means feed or supply) 16:47:18 a better Icelandic name for Forth would be Tvístakkur or some such of that effect. 16:48:33 --- quit: dys (Remote host closed the connection) 16:52:31 <`presiden> F012T 16:53:59 Forð 16:55:00 (ð is the IPA-chracter for the "th"-sound) 17:00:52 þe correct usage of þorn and eth is a bit tricky but wið a bit of practice it becomes easier 17:02:52 zún jú will bí able tú wræt English wið Icelandic-ish orðograffí 17:05:43 þis lúks a bit streinge tú natív English spíkers but is hkvæt servisable 17:08:36 okay, how many tried to plug the three lines above into Google Translate and what did you get? 17:18:02 --- join: mark4 (~mark4@172.56.7.6) joined #forth 17:19:58 --- join: mark4_ (~mark4@rrcs-67-79-14-26.sw.biz.rr.com) joined #forth 17:22:33 --- quit: mark4 (Ping timeout: 258 seconds) 17:27:56 --- quit: mark4_ (Remote host closed the connection) 17:50:02 --- join: X-Scale` (~ARM@180.61.158.5.rev.vodafone.pt) joined #forth 17:51:46 --- quit: X-Scale (Ping timeout: 244 seconds) 17:51:46 --- nick: X-Scale` -> X-Scale 18:25:06 --- join: tabemann (~tabemann@rrcs-162-155-170-75.central.biz.rr.com) joined #forth 18:27:21 --- quit: dave9 (Remote host closed the connection) 18:27:55 --- quit: dave0 (Ping timeout: 245 seconds) 18:28:48 --- join: dave0 (~dave0@069.d.003.ncl.iprimus.net.au) joined #forth 18:31:50 --- quit: deesix (Ping timeout: 246 seconds) 18:32:09 --- join: deesix (~dddddd@unaffiliated/dddddd) joined #forth 18:32:32 --- join: dave9 (~dave@069.d.003.ncl.iprimus.net.au) joined #forth 18:34:48 Why would anybody want to use Icelandic orthography when they could use Hangul? 18:36:22 wㅎy 워울ㄷ 이 완ㅌ 토 우세 이체란딫 오ㄹㅌ혹랖ㅎy w헨 이 초울ㄷ 우세 한굴? 18:36:27 Hangul isn't as simple and straightforward as one might like to think 18:36:33 Hangul looks way cooler. 18:37:48 in Iceland everything is cooler 18:38:06 Iceland is one of those inherently cool places 18:38:47 * tabemann modified hashforth so that it has four different sizes of literal, to allow more compact code 18:39:45 (i.e. why use 8 bytes when for a value from -128 to 127 a single byte would do) 18:39:56 for me, alignment of code addresses :P 18:40:36 in my forth I've basically decided to say to hell with alignment 18:42:25 I never had any alignment problems with my Motorola MC14500B code :) 18:43:15 then again a one bit cpu isnt a lot of fun ;-) 18:43:56 I had to look that one up 18:44:13 that's like the ultimate in bit-slicing 18:45:39 It was very popular in early traffic light controllers in Australia in the 70's 18:47:20 I made up a simple dev system for one around 78' including a membrane keyboard that was horrible to use so I never did make anything with it 18:47:56 what was horrible - the dev system or the keyboard? 18:47:58 and of course all those cheap slow hand calculators ... they were one bit also 18:48:30 the keyboard was horrible, the dev system was just the usual electronics for such a device 18:49:12 hand held calculators I mean 18:50:39 how does one even use a 1-bit microprocessor? 18:50:45 serially 18:51:10 ttmrichter: Yeah Hangul is a kind of orthography that uses smaller simpler symbols for sound and composes them into syllable symbols using simple rules. So if one comes across a syllabillic symbol one has not seen before one can read the sounds out of it pretty easily 18:52:21 tp so basically to add two 8-bit numbers one needs to issue 8 add instructions in a row? 18:52:29 tabemann: yeah, I am pretty glad to elide the current heatwave in Europe 18:52:29 that seems rather inefficient in space 18:53:20 tabemann, yes and dont forget the carry :) 18:54:27 it seems like shrinking beyond a certain point has diminishing returns 18:55:06 because like in this case one needs eight instructions to do the work of one on an 8-bit microprocessor 18:55:39 tabemann, it's a old cmos cpu, there wasnt many around in those days 18:56:42 it came before many of the old 8 bit cpus we all know today 18:56:55 even the 8008? 18:57:25 i dont know, one would have to search for the timelines 18:57:47 I've no doubt that it's not used anywhere for decades now 18:58:35 I bet there's some factory somewhere that still has it in its ladder logic 18:59:25 one never knows, the chip was very reliable 19:00:42 the wiki says that you can still find implementations on ebay today 19:01:47 wow 19:02:28 but who would want one, I love my cheap, fast, feature packed, low power stm32 19:03:46 I'm anxiously awaiting my STM32L4 board 19:04:35 ordered already ? 19:04:46 a discovery board or a nucleo ? 19:04:48 I'm divided between trying to write a full Forth implementation for it or whether I should write a tethered implementation 19:04:59 yep, I ordered a discovery board 19:05:24 theyre a bit better made I think 19:05:58 either way I'll be using the same runtime for each; the difference is in the runtime I'd load 19:06:03 if youre writing a tethered Forth then the stm32F0 would be a good choice as it can have as little as 16KB flash 19:06:33 but a tethered Forth for the STM32L4 probably makes no sense as it has TONS of Flash ? 19:07:02 for instance the Forth I use, Mecrisp-Stellaris only requires 19KB of flash 19:08:22 * tabemann wants to handle loading code the way Mecrisp-Stellaris does it, and is trying to figure out how to implement that 19:09:13 tabemann, I meant to mention, if you design for the STM32L4 it would also be good to have your Forth run on the STM32F0 family if your binary is small enough, but the F0 code while being upwardly compatible, has a few differences to M3 code 19:09:16 probably have a dedicated buffer into which code is to be written separate from data, and then when loading into Flash run a relocator against the data in that 19:12:31 tabemann, going to do any RE against Mecrisp-Stellaris ? 19:12:57 tabemann, Mecrisp-Stellaris is written in pure GNU assembly 19:13:35 stm32 is what kind of architecture again? 19:13:47 tabemann, I have found 'Cutter' pretty good for RE analysis 19:13:58 Zarutian, CortexM 19:14:19 ARM CortexM? 19:14:43 Zarutian, yes 19:14:47 There's another Cortex-M? 19:14:54 back 19:15:10 RE? 19:15:20 reverse engineering 19:16:47 I wasn't planning on reverse engineering per se; Mecrisp-Stellaris and hashforth are rather different in that Mecrisp is native-code compiling while hashforth is TTC, even though I might write a JIT for hashforth at some point 19:17:33 ttmrichter: contrary to what ARM wants you to beleave yes there was another Cortex. An old discontinued MCU architecture for low performance and low voltage. A failed Chinese flop of a thing. 19:18:10 I'm not sure whether I should try to compile (the currently POSIX) hashforth against an STM32 or whether I should rewrite the runtime from scratch in assembly 19:18:58 tabemann: how much of the runtime is POSIX specific? 19:19:34 Zarutian: currently there's a good amount of stuff that is POSIX-specific, but all that can be #ifdef'ed out 19:20:18 the only problem is I'd still need assembly to trap processor exceptions, so I might as well write the rest in assembly too 19:21:34 * tabemann currently traps exceptions with sigaction() and siglongjmp() 19:22:08 tabemann, I meant RE more in regard to looking at how Mecrisp-Stellaris handles the Flash side of things 19:22:43 Zarutian: There's Cortex (not ARM), and then there's Cortex-M, Cortex-A and ... Cortex-R, was it? 19:22:59 I'm pretty sure the Cortex-* stuff is always ARM. 19:23:59 I think it's Cortex-M, Cortex-A, and Cortex-R, and they're all ARM 19:25:29 my question is what is Cortex-R for 19:25:47 arm has a nice table showing the types somewhere, I saw it the other day 19:26:23 like it's obvious obvious what Cortex-M and Cortex-A are for - Cortex-M is for little, low-power, MMU-less systems and Cortex-A is for big, power-hungry, MMU-ful systems 19:26:35 never understood why many of these core designer companies insist of complexifing their cores out of the wazoo 19:26:35 true 19:27:44 tabemann: depends what you call an MMU but I suspect you mean at least virtual address translation and trapping on access violations. 19:28:01 The 'A' Application profile features powerful processors found in high-end products like smartphones, tablets, or netbooks. This includes the famous Cortex-A8 and Cortex-A9 (in your super phone) processors. 19:28:02 The 'R' Real-time profile processors can be found for example in control units for automotive systems or hard disk drive controllers. They come with specific features suited to real-time environment constraints. R profile processors have for example a tightly coupled memory,local to the processor, that allows for fast-responding code and data. They also all have hardware support for integer division, that would otherwise require software emulation. 19:28:02 The 'M' Micro-controller profile processors are smaller and used in numerous embedded systems like human interface devices, automotive control systems, power management systems, and others. 19:28:16 I have seen segmentation via two boundary registers used. 19:28:41 Zarutian: yes that, not more archaic systems with base and bound registers 19:30:17 tabemann: no not the old base and bound registers. There were literally two boundary registers spliting address space into read-only&exec, read-only and read&write segments, in that order. 19:31:51 I am fine with much simpler MMUs like was used on earlier PowerPC architectures, where an TLB miss caused an trap to kernel handler. No expensive and inflexible pagetable walking hardware. 19:33:07 that to me still is a sort of MMU, albeit a more limited sort 19:33:51 heck read an reference to an architecture where priviledged code (code running in kernel mode) could read out the current TLB. Made context switching 'rewarmup'-less 19:34:06 I read* 19:35:18 but I prefer the Transputer or Green Array approach of many simple cores having their own dedicated working and program memory each 19:35:57 in my current Forth I'm just relying upon the MMU and the OS to handle paging - I have code written to trap segfaults, and to dump a backtrace and recover from them 19:38:39 tabemann: so not a bare metal forth then 19:38:52 I'd need to rework that when I try to get it to run on bare metal 19:39:36 I'm going to have to rework my whole signal/interrupt handling system 19:40:54 Zarutian, do you own a GA-144 by any chance ? 19:41:22 tp: nope. 19:41:50 same 19:42:30 ttmrichter, have you written about the GA-144, any opinions on it's use ? 19:43:02 tp: personally I think Green Arrays the company does not know what they are doing business wise 19:43:38 okay, I'm gonna head home now - bbl 19:43:41 * tp has absolutely no idea what one would use a GA-144 for, but thats only a reflection on my lack of intelligence 19:43:57 cya tabemann 19:44:55 Zarutian, it's well known that Chuck More isn't a business man, and has been previously ripped off by a business partner, but hes a great Forth mind :) 19:46:00 how many companies have utterly failed in the marketplace even tho they had the 'best' product ? and how many times was the reverse true ? 19:46:26 marketing and shark like salesdroids seems to win every time 19:46:42 tp: yes hence why GA is so mistrustfull of biz people. Personally, anyone with only an MBA must be superwised. 19:48:29 it seems the Risc-V has generated far greater excitement than the GA-144, even amongst Forth addicts 19:48:36 --- quit: tabemann (Ping timeout: 272 seconds) 19:49:07 regarding 'shark like salesdroids': I know of an company that sacked a gullible middle manager that bought into the crap a salesdroid fed hi,. 19:49:20 s/hi,/him/ 19:49:24 is it a case that none of us are clever enough to see what Chuck sees, or perhaps technology and money issues resulted in the GA-144 being handicapped in some way ? 19:50:02 Zarutian, wouldnt be the first time a salesdroid conned management 19:50:54 the best part when the company with the salesdroid tried to deliver or collect on an invoice? They were directed to that schmuck 19:52:23 god save us from salesdroids 19:53:18 specially those who make up features on the spot EVEN though they have been given clear instructions not to go outside the product brief 19:53:58 Western Digital has already included the RISC-V core in it's hard drive products, but I dont know of one product that uses the GA-144 ? 19:54:26 salesdroids! ... shudder 19:55:10 GA-144 is a bit too wierd for most to know how to even attempt at program. 19:55:45 I have only heard of it being used in low power DSP like work and not much detail about that 19:55:59 is thats because the GA-144 is too far ahead of it's time, or just a inventors pet project ? 19:56:30 something like FlowBasedProgramming might help and each 'computer' having a bit more memory to work with. 19:56:42 a bit of both 19:57:38 personally I'm still working on understanding all the peripherals in a stm32f051 which is about 7 years old now, and I have absolutely zero clue where to start even trying to see where a GA-144 would excel 19:57:53 plus GA-144 suffers from being only micron feature size tech while most stuff nowdays are at least 250 nm feature size. 19:58:44 similar to the risc-V as the old 900nm fabs have all repaid their investments by now and offer cheap fabbing 19:58:50 tp: those peripherals, are they accessed via memory address space or? 19:59:30 the stm32f051 (cortex-m0) are addressed via memory space yes 19:59:59 simple memory mapping gets to them 20:00:17 hell, if an highschool student can do 500nm (half-micron) feature size with mish-mash of donated old fabber tech then doing RISC-V should be breezy 20:00:19 tp: I tried to get something with the GA-144 going but found Greenarrays impossible to deal with. 20:00:24 $40026000 constant AES ( Advanced encryption standard hardware accelerator ) 20:00:25 AES $0 + constant AES-CR ( control register ) 20:00:31 ( example) 20:00:53 ttmrichter, damn, youd be the one to see the possibilities! 20:00:53 I've given up on doing anything with their kit as a result. I like the GA-144 concept but am unlikely to ever be able to use it for anything. 20:01:40 I LOVE the idea. 20:01:49 But it's not much use if you can't source the parts for prototypes. 20:02:09 Well, source the parts *AND* get the technical assistance required to make them work. 20:02:37 I know the GA-144 chip is $20 in a pack of 10 ? but the development board is very expensive and perhaps inhibits its sales ? 20:02:59 this is where Texas Instruments shines. They make a new chip, make the datasheet and a bringup howto available and sell the chip at cost. 20:03:19 That was my first pitch to them, actually: a development board that isn't stupidly expensive *AND* that provides an architecture that doesn't require as much magic. 20:03:38 They were even receptive to the idea. 20:04:00 ttmrichter: just fuddy duddy in getting on with it? 20:04:36 But ... wow. Six week turnarounds on emails. Not one, not two, but THREE Skype call appointments with their "sales engineer" that were ghosted, and then a scolding when I asked for a time frame on actually getting this? 20:04:38 I guess that GA-144 development board is the proof of the business failings so far, perhaps chuck just doesnt care ? 20:04:46 I dropped the project idea and haven't looked back. 20:05:20 I was getting the impression that Greenarrays has one or two very large 900-pound gorilla customers that occupy all their resources. 20:05:37 ttmrichter: holy shit man. Worse than many big lumbering companies I had unpleasentness to deal with. 20:06:00 yeah, it's not like the GA-144 can open a portal into another dimension or anything, and the STM32L0xxx chips are damn hard to beat economically 20:06:02 And that they don't understand the notion of "diversified customer portfolio" and 900-pound gorillas being destructive to corporations at an existential level. 20:06:05 oh, so they are basically toast if that gorilla disapears 20:06:31 but who is the gorilla ? 20:06:33 Absolutely. 20:06:43 Sounded governmental from the little clues in the conversation. 20:07:36 as we are on the topic of component/subsystem sourcing, one thing I absolutely hate is 'call for quote' on product listings. 20:07:55 perhaps the US govt uses the GA-144 inside all their millions of surveillance cams for face recognition processing? 20:08:26 Zarutian: "call for quote" is code, for me, meaning "I don't want your business". 20:08:34 "I am still at the feasibility prototype stage man and I want to see if the product I have in mind could actually work and you want me to get a quote from you?" 20:08:36 same here 20:08:48 When I design, I need to know how much parts are going to cost me. If you can't give me a stable baseline, fuck you. 20:08:59 typical sales droids in action 20:09:55 ttmrichter: indeed. And I have seen less and less 'call for quote' on sites like digikey. I suspect that they (the distributor) are seeing that 'price point' being excluded in parametric searches. 20:09:56 in Australia shipping has long been that type of deal. Tell us how many boxes you ship a day and we *might* quote you for a shipping account 20:11:00 luckily Fedex has taken over our biggest shipper and now there is a easy to use website, no account needed, enter locations etc and get a instant costing 20:11:24 I keep getting gob-smacked by just how UTTERLY SHITTY logistics are in the west these days. 20:11:40 On Sunday I'm about to embark on a round of purchases (personal, not corporate). 20:11:41 all these "call for quote" systems are ancient and I think they are in line to be taken over by smarter players 20:12:01 what many of these salesdroid dont understand that many companies buy a few components for R&D through shells or cutouts to see if those components work out. 20:12:03 I'll wind up ordering about 20 or so different packages from 20 different cities across China. 20:12:35 "but we dont want to deal with small fish customers." dude, some of those can actually be or become huge customers of yours. 20:12:35 ttmrichter, hope it isnt during a Chimese holiday season ;-) 20:12:38 Connection to these companies is through one of the world's largest B2C and C2C networks ever. 20:13:09 Goods will vary in kind from tea to foods to consumer electronics to dev components to ... playing cards. 20:13:15 (Yes. All in my purchase queue.) 20:13:17 Zarutian, it's been like this for ever in electronics ... sadly 20:13:36 I have a single app that will track all the packages. 20:13:51 tp: was forever but frankly those 'call for quote' companies are either dieing or adapting. 20:14:13 At any given point I can find out where any of my packages are in the country within about a four-hour window. 20:14:28 ttmrichter: wait when you say shitty logistics in the west you mean USA only or? 20:14:35 There'll be a dozen different logistics companies in the mix, but a single app covers them all. 20:15:03 Right down to mapping where the package was last seen with details like the number of the truck. 20:15:37 remember when MOS tech held a 6502 event in the USA and sold them from cookie jars for $15 each or something ? At the time motorola had the awesome 6800 which was like $40 and you had to order thousands or they wern't even interested ? 20:15:43 Zarutian: USA, Canada, most of Europe (when I compare notes with some of my Belgian and French friends), Australia (culturally western if not geographically), ... 20:15:57 as a result the dam 6502's went into everything 20:16:01 NOTHING I've even heard rumours of matches the logistics of purchasing here. NOTHING. 20:16:43 ttmrichter, yet outside china when I buy something it comes via 'china post' and takes 3 months with NO tracking ? 20:16:43 I was shocked at how low-tech and primitive shopping was in Canada when I visited there a couple of years back. 20:16:58 tp: Two reasons for that: 20:17:14 is it a case of the Chinese attitude of 'fuck outsiders" ? 20:18:01 ttmrichter: I can beleave it about USA and Canada and some of the old-style Europe countries but not the nordic countries. Heck I even managed to reroute a delivery to where I was on that day when the package was being driven out to me. 20:18:10 tp: 1. They're exploiting a loophole in international postage agreements that makes posted stuff essentially free shipping. 20:18:44 tp: aah right, you are doing that kind of shipping via post and not cargo, yes? 20:18:52 2. Chinese logistics companies are all domestic-only. 20:19:05 Summary re Greenarrays: If ttmrichter LOVES the idea of the GA-144 but ran into a brick wall dealing with them, then thats all I kneed to know as to why the chip isnt exactly popular 20:19:32 Zarutian: do you have any orders outstanding? 20:19:51 If so, can you tell the exact delivery status from point of sale to home at any time? 20:20:00 ttmrichter, thanks for explaining that 20:20:00 Across multiple businesses and carriers? 20:20:04 ttmrichter: no, not personally. I dont buy much stuff. 20:20:06 Single point of access? 20:21:17 In Canada someone was telling me that they could track any shipment any tme. 20:21:19 it isnt on that level yet of Single point of access but I have gotten the tracking code from where the package was handed over to the next carrier. 20:21:45 They just needed to know the tracking number and the precise delivery company, visit that company's web site (or app on phone), enter that tracking number manually, ... 20:22:08 --- quit: dave0 (Quit: dave's not here) 20:22:18 yeah, sounds like those Taxi apps that dont get why uber or lyft is so popular 20:22:35 When this next round of orders starts, about a day in I'll do a screen-capture video of how I track packages. 20:22:46 please do. 20:22:49 It's mind-boggling how integrated all these systems are. 20:24:03 regarding post vs cargo: the former is covered by all sort of exceptions and rules set by the Universal Postal Union while the latter usually follows a simple formulae 20:25:03 Oh, more fun. 20:25:04 geodesic distance times weight times size-ackwardness plus handling-ackwardness kind of formula 20:25:14 Package delivery day arrives. 20:25:17 I'm not at home. 20:25:35 So it goes to any number of large delivery boxes that litter my compound. 20:25:50 compound? 20:26:21 Residential compound. Gated community would be the term, only it doesn't have the stigma of snobbishness here. ALL communities here are gated communities, in effect. 20:26:36 My purchasing app gets the code. As does my tracking app. And I get the code delivered by SMS as well. 20:27:11 I go to the box and enter the code. Or just open my tracking app and click on the package notification. The right door opens and I get my stuff. 20:27:11 sorry, I just had a wierd association regarding the word compound. Only ever heard it used in military speak. 20:27:42 where is this? China? if so which province? 20:29:02 China, Hubei, Wuhan, Hong Shan 20:29:11 ttmrichter, I guess Canada won't ever be tempting you back home ? 20:29:34 yeah there is something similiar here in Iceland called Póst-Box but there isnt for compounds but at hub areas. 20:29:40 Not in the near future, no. Maybe in the distant. And my son will probably be going there for middle school. 20:30:30 Zarutian: These are independent of the post office. (Though China Post upped its game tremendously in the face of competition from aggressive logistics outfits too. They deliver to my home. On Sundays. At 10PM if necessary.) 20:30:32 * tp lives in a 'gated community' if a large shed with a surrounding barbed wire security fence and nightly locked gates applies to that term ;-) 20:33:18 ttmrichter, logistics in China sounds incredible, but I imagine large spaces between rural towns with hills and poor electronics communications, is this not so ? 20:33:59 also narrow rocky roads blocked with slow bullock pulled carts ? 20:35:00 hmm Australia is probably more like that scenario than china except we have few bullocks and no carts 20:35:36 ttmrichter: what are the payment options for this kind of purchasing? still only credit cards? or have they progressed beyond that? 20:36:40 What's a credit card? 20:36:58 I haven't had a credit card for 18 years almost. Coincidentally I've been in China for 18 years... 20:37:18 The single largest payment options for this in China are AliPay or WeChat Wallet. 20:37:34 And these are used at EVERY LEVEL imaginable. 20:38:02 And Alibaba are their own bank ? 20:38:17 I haven't used cash for ANYTHING in months and I haven't used it for anything sizable in years. 20:38:21 people pay alibaba directly, no bank to go thru ? 20:38:34 I have started to see immiations of this kind of thing here. 20:38:35 My Ali account is backed by a bank account in theory, but most times I just keep a balance in Ali. 20:38:37 i was reading about this the other day 20:38:55 and no Ali 'account keeping fees' ? 20:39:03 (like every bank) 20:39:36 Síminn Pei (which only a handfull of places accept) and other wannabes or imports like Apple Pay (only a few merchants have bothered setting up for those) 20:40:22 tp: what kind of a bank account has 'account keeping fees'? 20:40:58 Most American and Canadian banks. 20:41:04 They call them "service fees" but ... 20:41:09 * Zarutian still has a few bank accounts with less than 30 kr that he has been trying to close for years but no fees for keeping them. 20:41:14 The "service" is often "account exists". 20:41:48 that is so damn bizzare to me. 20:42:23 Banks found out that they can do literally anything to customers and nobody will raise a stink. 20:42:41 Government will look the other way and customers will whine, but not do anything like, say, changing service providers. 20:43:17 (Not that changing does any good given that the number of actual competitors is an order of magnitude fewer in Canada and TWO orders of magnitude fewer in the USA than it was 25 years ago.) 20:43:36 oh that explains why some companies here and other places have policies like "Bank Of America Cheques ARE NOT ACCEPTABLE PAYMENT METHOD" 20:44:32 say, does AliPay or WeChat Wallet accept topups via cybercoin (or cryptocurrencies)? 20:45:25 * Zarutian has found bank wires utterly expensive and unreliable due to 'omg terrorism' "derisking" crap or some such 20:46:48 regarding USA and Canadian banks: so this is why some younger peeps I know there are slowly going unbanked or, what was the term, minbanked 20:51:09 --- join: tabemann (~tabemann@2600:1700:7990:24e0:1ce2:9bd6:6ee:b28e) joined #forth 20:53:02 --- part: tabemann left #forth 20:53:54 --- join: tabemann (~tabemann@2600:1700:7990:24e0:1ce2:9bd6:6ee:b28e) joined #forth 20:54:44 --- join: gravicappa (~gravicapp@h109-187-47-207.dyn.bashtel.ru) joined #forth 21:28:35 --- quit: dddddd (Read error: Connection reset by peer) 21:48:44 ttmrichter, is so right, charging fees for the existence of a account is one way they make money 21:55:32 Zarutian: Yep. If I were in Canada I'd have the minimal bank account possible and do everything cash. 22:01:01 I switched my order for the Discovery board from some random distributor to ST itself 22:01:05 it turns out it's much cheaper 22:01:06 but 22:01:17 ST is fucking intrusive in wanting to know what you plan on doing with their board 22:01:52 it's like if you ordered a machine from Dell and Dell wanted to know whether you plan on watching porn with it 22:01:56 i wouldnt worry about it, it's just some form that a sales droid wrote and no one will ever read it 22:02:13 --- quit: proteusguy (Remote host closed the connection) 22:02:30 also, when I made my order, it then said it was "Under Review" 22:02:46 --- join: proteusguy (~proteusgu@cm-58-10-208-146.revip7.asianet.co.th) joined #forth 22:02:47 --- mode: ChanServ set +v proteusguy 22:02:57 that's the first time I've ordered something and they've come back and said it was "Under Review", as if they might let me buy it 22:03:14 the data will be mixed in with millions of others by a machine and end up on a power point chart at the sales droid monthly meetings 22:04:18 they have to conform to 'supersecret American technology export laws' so thats probably what they mean there 22:04:50 it's a fucking Discovery board 22:05:07 it's not like it's going to be used in some Chinese supercomputer to do calculations on Chinese nukes 22:05:20 yeah but it could be used to power a north korean ballistic missile! 22:06:07 (the US actually forced Intel to not sell Xeon Phi's to the Chinese for use in supercomputers for ostensibly that reason) 22:06:33 tabemann, youre right, everyone know the Chinese supercomputers are made up of millions of racks each containing hundreds of ESP2866's 22:06:53 lol 22:08:47 --- join: dave0 (~dave0@069.d.003.ncl.iprimus.net.au) joined #forth 22:09:00 tabemann, when I order from arrow.com (usa) I get a similar disclaimer that they *might* not be able to sell it to me, but so far so good 22:09:08 re 22:09:27 dave0, g;day cobber 22:09:34 hey tp 22:09:51 hey dave0 22:10:18 hi tabemann 22:10:21 I wonder how when they asked me the name of my project I wrote "Foo" will go over 22:10:25 Yipee!! I pot in a request for Pictured Numerical Output to be added to Mecrisp-Across and it's just been released :) 22:10:56 tabemann, they may think youre making a alien "foo fighter" ? 22:11:14 well on the next line I wrote "Bar" 22:11:56 they may think it's for use in a alcohol drinking establishment ? 22:13:53 another thing is that they forced me to register for their site, and in doing so they required me to name my employer, and they are putting it on the shipping label even though I'm not purchasing this for my employer - as if they can't imagine someone buying one of these boards all for their own purposes 22:14:54 sales droids ... 22:15:29 I've done all this with them, never had any issues, that info goes into a big black hole 22:15:51 I get that 22:16:09 but part of me is like "it's a fucking discovery board - that's not something companies buy" 22:16:10 I do get irregular emails from ST telling me about their latest chips, which I like to read 22:16:41 it's a 'product', thats all the sales droids understand 22:17:10 part of me wants to go and cancel my st order and go and buy it off amazon again 22:17:43 I bet billions of STM32's found their way into products because a engineer bout a disco board once ? 22:18:42 disco boards are that first dose of STM32 microcontrollers 22:19:23 plus a ST account is handy when you want files such as SVD's 22:20:22 I initially bought six STM32F0 Discovery Boards from AVENET I think it was 22:20:43 back in 2014 22:21:15 lately I have bought chips from arrow.com while they had free shipping to Australia ! 22:21:31 tabemann, do you have a arrow.com account ? 22:21:44 no I do not 22:22:26 as youre in the USA it's probably not as beneficial as Australia because I think arrow.com have free shipping in the USA aNyway 22:23:00 their stm parts are genuine STM and they have some excellent 'overstock' prices 22:23:22 they give 10% discount on your first (or second) order as well 22:24:08 I bought ten STM32L162RD chips for $9.00 aud inc free shipping that way 22:24:20 and the STM32L162RD is insane! 22:29:58 now the hard question 22:30:16 wtf do I do with this board other than blink LEDs and display stuff on the segmented display 22:31:55 use it to test a hashforth ? 22:33:29 I plan on doing a full rewrite of hashforth for it 22:33:35 write an ARM assembler in forth 22:33:51 and use that to write hashforth-arm in ARM assembly 22:34:26 this is all so I don't have to figure out how to do gcc crosscompilation and figure out how to do inline assembly with gcc 22:34:47 and also so I have an excuse to write hashforth in assembly 22:34:57 plus it will run Mecrisp-Stellaris 22:36:19 there are 11 Forth programs for that board included in the tarball 22:37:43 I'll have to figure out how much I will have to cut down the hashforth "userspace" to get it to run on the discovery board 22:38:05 but I don't anticipate I will have to cut down the code by much with that awesome 1 MB flash memory 22:38:37 the only question is that hashforth as currently designed allocates a lot of memory to things like line editor history 22:38:45 and that would have to be cut down severely 22:39:07 since the board only has 128 KB of RAM 22:39:16 *only* 22:39:38 also I will have to take out the heap 22:39:38 I have 8KB on my stm32f051 chips 22:39:40 --- join: dys (~dys@tmo-123-188.customers.d1-online.com) joined #forth 22:39:46 because the heap takes a lot of memory space 22:40:00 but hashforth is modular enough that that can be done 22:40:24 since the only thing which the heap is necessarily used for is loading code from files 22:40:36 and that won't be an issue with hashforth-arm 22:41:53 one thing I will have to change is how hashforth boots 22:43:07 currently hashforth receives a big image, containing both hashforth VM code, which it relocates, and uncompiled hashforth code, which the relocated VM code executes and which generates new hashforth VM code 22:43:14 this happens every time 22:43:31 what would need to change is this would have to be a one-time process 22:43:39 when going from resource unlimited PC's to limited microcontrollers I think there would be a multitude of challenges 22:44:39 which after receiving the image it would relocate it, and it woule compile the uncompile hashforth code, which would have to include calls to transfer the compiled code into flash 22:44:55 and after that, with subsequent boots, all it would do is set up the variables in RAM that were lost 22:46:36 there is one significant architectural change that I'd have to deal with 22:46:44 hashforth is a TTC forth 22:47:06 meaning there is a central table containing information on all the possible words - and this table is fixed sized 22:47:39 for a microcontroller hashforth there isn't enough space in the RAM, even with 128 KB, to store a significant sized table 22:47:49 so it has to go into flash 22:48:04 but we can't just be updating flash every time we compile a word 22:48:10 so the table will have to be split 22:48:15 with one big table in flash 22:48:20 and one little table in RAM 22:48:56 what will happen with the flash table is it will contain dummy entries which will be overwritten as new entries are compiled into it 22:48:58 aha 22:49:12 and these dummy entries will mean "nothing here, look at the RAM table" 22:49:29 it will be a performance hit, but only for code that has not been relocated into flash yet 22:49:35 Mecrisp-Stellaris can compile Words into ram or flash on a word by word basis 22:50:22 I'm always flashing and erasing Flash 22:50:36 that's hard for me to do, because tokens are assigned upon compilation, and linearly increase 22:50:48 what I can to is modify MARKER 22:51:01 to reflash flash and remove entries from the flash token table 22:53:40 sounds like a major redesign 22:54:31 now that I've been thinking about it's simpler than I thought 22:55:08 what I really need is to write a relocator - code to take hashforth VM code, and move it in memory, on the fly 22:55:15 it's doable, definitely 22:55:55 I also need a means to tell the VM that a word has moved - but that is pretty simple 22:57:04 about MARKER, that is just the current MARKER combined with reflashing the flash to overwrite the erased flash token table entries with dummy values 22:59:55 all in all this design is pretty simple and workable, its only real limitation is that one cannot compile individual words to flash or to RAM, rather all words up to a certain point must be transferred from RAM to flash 23:00:47 hmm 23:01:32 how I develop is I have a bunch of working words I load into flash. I then test new words in ram and when theyre good, I move them to flash also 23:02:39 that way I can do my usual 100 million resets of the MCU to clear ram while developing without needing to reload heaps of working Words 23:04:07 of course every embedded Forth user has their own system of development 23:05:50 --- quit: dys (Ping timeout: 245 seconds) 23:06:09 well, actually, now that I think of it, I could do it how mecrisp-stellaris does it 23:06:27 by allowing "holes" in the flash token table 23:06:38 and also not assigning tokens in order 23:06:47 but by the first vacant location 23:09:47 the only problem is 23:09:47 --- quit: `presiden (Read error: Connection reset by peer) 23:09:53 to make the most efficient use of RAM 23:10:05 the whole token table would have to be shifted over 23:10:14 the RAM token table that is 23:10:27 when its lowest-index token is pushed to flash 23:10:40 and likewise 23:10:50 in the compilation buffer 23:10:52 in RAM 23:11:00 to avoid bubbles 23:11:21 it would have to be shifted over with relocation 23:11:46 whereas simply pushing all the currently-compiled words in RAM out to flash is much simpler 23:33:36 --- join: proteus-guy (~proteusgu@cm-58-10-208-146.revip7.asianet.co.th) joined #forth 23:59:59 --- log: ended forth/19.06.28