00:00:00 --- log: started forth/19.04.28
00:04:18 --- quit: dave0 (Quit: dave's not here)
01:23:47 --- quit: rain1 (Ping timeout: 250 seconds)
02:02:54 --- quit: ashirase (Ping timeout: 245 seconds)
02:07:24 --- join: ashirase (~ashirase@modemcable098.166-22-96.mc.videotron.ca) joined #forth
02:33:31 --- join: rain1 (~My_user_n@unaffiliated/rain1) joined #forth
03:50:03 --- join: dddddd (~dddddd@unaffiliated/dddddd) joined #forth
04:08:50 --- join: rdrop-exit (~markwilli@112.201.166.63) joined #forth
04:19:51 <WilhelmVonWeiner> been playing with graph databases this morning
04:20:06 <WilhelmVonWeiner> really cool stuff
04:22:32 <rdrop-exit> I'm watching today's SVFIG meeting on Youtube
04:22:48 <WilhelmVonWeiner> what are the talks about today
04:23:34 <rdrop-exit> Right now it's Intel 4004 related
04:23:42 <proteusguy> the only real actual forth thing was pointfree 's chat about Forth unikernels which actually is something I'm quite interested in.
04:24:20 <rdrop-exit> Haven't gotten that far in the video.
04:24:49 <proteusguy> rdrop-exit, I skipped ahead and now went back to the Intel 4004 stuff. 
04:29:01 <presiden> SVFIG?
04:29:18 <proteusguy> presiden, yep https://www.youtube.com/watch?v=4I74wNEDMEY
04:30:53 <rdrop-exit> I'm mostly listening rather than watching, have it running in the background
04:34:19 <presiden> interesting
06:43:39 --- quit: rdrop-exit (Quit: Lost terminal)
07:13:52 <siraben> proteusguy:  forth unikernels?
07:14:07 <siraben> I'm not familiar with the concept.
07:16:00 <proteusguy> effectively you create a combined executable and OS all in one that doesn't have the overhead and complexity of something like linux or windows just to run your forth app. sorta like self-hosted forth except you are using existing device drivers and such.
07:16:51 <proteusguy> https://en.wikipedia.org/wiki/Rump_kernel    http://rumpkernel.org/
07:16:58 <PoppaVic> cp/m - dos
07:17:10 <siraben> Interesting.  That is definitely achievable with zkeme80
07:17:28 <proteusguy> I think you're already there with zkeme80. :-) 
07:18:11 <siraben> I'm trying to weigh in the upsides and downsides of adding threads with a round-robin interrupt scheduler
07:18:16 <siraben> Isolating Forth "processes" with their own stacks is quite difficult
07:18:26 <siraben> I mean their own heaps
07:18:39 <siraben> The stacks are somewhat easy because Z80 allows direct manipulation of the stack pointer
07:18:59 <proteusguy> classically I think systems just divided up the memory in a fixed manner.
07:20:58 <siraben> But realistically I think that full-blown "concurrency" isn't really needed (especially on embedded, calculator-like systems)
07:21:17 <siraben> Persistence I guess is more important, I should be able to restore an application's state fully, for instance
07:21:30 <siraben> Doesn't require the complexity of threads and so on
07:22:10 <proteusguy> eventually we got chips that provided memory management and you could map local "virtual" memory to different physical allocations. Later z80s and 68881 (as I recall from memory) started offering this in hardware as the 64k limit was blown thru.
07:22:32 <siraben> Yep. The TI-84+ uses such a scheme
07:22:36 <siraben> Talking to flash chips
07:22:44 <siraben> IIRC 24 bit virtual memory space
07:23:36 <siraben> Without all the bloat of the default operating system it's actually an incredible amount of space one can play with, I've barely scratched the surface of it.
07:24:34 <siraben> I think the next big leap would be when I figure out how to send/receive via the USB port, so the entire operating system doesn't need to be reflashed every time I want to make a change to a single Forth module
07:28:17 <MrMobius> siraben, how does the 24bit pointer work?
07:28:34 <MrMobius> is it like three byte wide registers?
07:31:09 <siraben> MrMobius:  details here: http://wikiti.brandonw.net/index.php?title=83Plus:Memory_Mapping
07:31:10 <proteusguy> I'm guessing two overlapping 16-bit segmented registers.
07:31:48 <siraben> Essentially, I write to a port to set which memory bank is reflected in ram
07:31:56 <siraben> http://wikiti.brandonw.net/index.php?title=83Plus:Ports:06
07:32:02 <siraben> Port 6, for instance, swaps pages into memory bank A
07:32:08 <siraben> 4000-7FFF hex
07:32:57 <siraben> There's a bunch of raw flash routines: http://wikiti.brandonw.net/index.php?title=83Plus:OS:Raw_Flash_Commands
07:33:01 <proteusguy> ah worse than I thought... haha intel wasn't so bad after all.
07:33:20 <siraben> As usual with flash, writing is easy, can be done at random, erasing is an all-or-nothing operation
07:33:32 <MrMobius> siraben, ah ok so not virtualmemory then just page banking
07:33:42 <siraben> Right
07:34:02 <siraben> You could set up a Forth system to hide this implementation detail
07:34:11 <siraben> Something like !! and @@ for 24-bit reads/writes
07:34:30 <siraben> Except that !! would only change a 1 to a 0
07:35:43 <siraben> The only forth word I have that exposes this bank switching thing is https://github.com/siraben/zkeme80/blob/fbc09809196fd4f21dccafcd859943fc3a87b674/src/forth.scm#L1986
07:36:16 <siraben> Oh there's a CMOVE-FLASH too, although I haven't tested it
07:36:44 <MrMobius> interesting
07:37:49 <siraben> Still amazed that this assembler works,
07:38:01 <siraben> Didn't do anything too fancy, hasn't warranted a rewrite at all
07:38:23 <siraben> And I get extremely powerful macros for free, because it's Scheme
07:38:27 <MrMobius> I was thinking about how to implement that kind of system where you have objects spread out over different pages and combine them into a third object
07:38:49 <MrMobius> like when you do math on the calculator
07:38:49 <siraben> MrMobius:  define "combine"
07:38:56 <MrMobius> z=x+y
07:39:23 <MrMobius> the easiest thing is to reset the bank once for each source and once for the dest
07:39:38 <MrMobius> even though that is unnecessarily slow if they all happen to be in the same bank
07:39:44 <siraben> Right
07:39:58 <siraben> So it would work if I had larger functions that needed to be on different blocks
07:40:22 <siraben> And the kernel would figure out to do its swapping etc, and compile that word so that subsequent calls are fast
07:40:43 <siraben> The system would have to be really large by then to warrant such complexity
07:42:02 <MrMobius> you mean optimize so that it doesnt set the bank if its already pointing to the right bank?
07:42:18 <siraben> That could be a possibility
07:42:29 <siraben> So predefine the word headers in the first bank, say
07:42:38 <siraben> With pointers to their actual location
07:42:44 <siraben> i.e. what byte of what bank
07:43:05 <siraben> So initially the system could be slow, with a lot of swapping, but this only happens on the first invocation of each function
07:43:41 <siraben> so if I wrote ln(sec(x)), the system would go and compile SEC and LN and run them
07:43:57 <siraben> Next time you call LN or SEC it's just found like a normal forth word
08:17:22 <pointfree> At svfig, John Rible mentioned a nifty forth spreadsheet software where you could open a spreadsheet cell and put forth code in it.
08:19:40 <pointfree> I really want something like that using the octopart api for optimizing pcb bill of materials.
08:22:31 <pointfree> Then I found  https://en.m.wikipedia.org/wiki/Logical_spreadsheet
08:23:13 --- quit: X-Scale (Ping timeout: 246 seconds)
08:25:19 --- part: PoppaVic left #forth
09:02:48 --- join: X-Scale (~ARM@31.22.201.96) joined #forth
09:55:01 <john_cephalopoda> pointfree: Heh, reminds me a bit of emacs, where you can make speadsheets in org-mode syntax and fill the cells with (elisp?) logic.
10:01:25 <pointfree> Cool! 
10:19:29 <john_cephalopoda> pointfree: Is there a link to that talk? I couldn't really find it.
10:21:49 <pointfree> john_cephalopoda: https://www.youtube.com/watch?v=4I74wNEDMEY&t=8970s
10:23:48 <john_cephalopoda> Thanks
10:24:59 <john_cephalopoda> Ow, google hangouts...
10:25:58 <pointfree> :)
10:27:48 <john_cephalopoda> The German Forth club uploads their videos after the conference, with pretty good audio and video quality.
10:31:57 <pointfree> Forth-Gesellschaft is enviable in a lot of ways. They even have their own magazine, a good way to improve my German skills.
10:33:01 <john_cephalopoda> They also recently updated their website from a early-2000-ish site to a modern one.
10:33:18 <john_cephalopoda> SVFIG and forth.org, in comparison, look horrible.
10:33:49 <pointfree> Why do the Germans like Forth so much?!?
10:34:29 <john_cephalopoda> Oh, wow, their latest magazine is practically all about fixed point numbers and calculations <3
10:40:27 <john_cephalopoda> Some of the articles are even in English
10:40:55 <pointfree> http://forth-ev.de/wiki/res/lib/exe/fetch.php/vd-archiv:4d2018-02.pdf#page=13 "Sechs Pixel gehen in eine Bar und bestellen ein Sixel . . ."
10:42:21 <pointfree> link for the lazy http://forth-ev.de/wiki/vd-archiv
12:19:34 --- quit: gravicappa (Ping timeout: 245 seconds)
12:25:46 <john_cephalopoda> Heh, I am using OSS emulation to play beeps with forth :D
12:39:57 --- join: gravicappa (~gravicapp@h109-187-32-71.dyn.bashtel.ru) joined #forth
13:06:13 --- quit: gravicappa (Ping timeout: 250 seconds)
13:35:48 --- join: tabemann_ (~tabemann@172-13-49-137.lightspeed.milwwi.sbcglobal.net) joined #forth
13:38:41 --- quit: tabemann (Ping timeout: 276 seconds)
16:05:52 --- join: ullbeking (sid5364@gateway/web/irccloud.com/x-zwmtaikpysdfbclw) joined #forth
16:06:02 <ullbeking> good evening all
16:23:03 --- join: dave0 (~dave0@108.060.dsl.syd.iprimus.net.au) joined #forth
16:24:40 <dave0> hi
16:32:20 <john_cephalopoda> Hello ullbeking and dave
16:32:43 <dave0> hi john_cephalopoda 
16:38:11 <ullbeking> hello john_cephalopoda 
16:38:29 <ullbeking> wow it has been a long long time since i used forth
17:01:29 --- quit: john_cephalopoda (Ping timeout: 276 seconds)
17:15:09 --- join: john_cephalopoda (~john@unaffiliated/john-cephalopoda/x-6407167) joined #forth
18:52:24 <pointfree> Hello ullbeking
19:13:42 --- join: PoppaVic (~PoppaVic@unaffiliated/poppavic) joined #forth
19:13:52 <ullbeking> hi pointfree 
19:15:49 <ullbeking> i initially popped in here because i've started trying to build up a library of open firnware based documentation, but then i remebered how much job i had using forth in the 1990's, and i'm wondering, escpially, what are some apps it's used in now..?
19:15:56 <ullbeking> i know it's used a lot in embedded stuff
19:16:20 <ullbeking> and i am taking a new interest in embedded computing, which is a nice coincidendence
19:17:40 <ullbeking> p.s. why is there one compiler for arm and one for x32 at https://github.com/mark4th?
19:18:45 <ullbeking> i'm coming from a ppc/power8+9 background, so will any forth tooling avaiable to my require ]changing?
19:26:24 <PoppaVic> yes
19:50:59 --- quit: MrMobius (Read error: Connection reset by peer)
20:04:54 --- join: MrMobius (~default@c-73-134-82-217.hsd1.va.comcast.net) joined #forth
20:05:02 --- quit: dddddd (Remote host closed the connection)
21:02:50 --- join: gravicappa (~gravicapp@h109-187-32-71.dyn.bashtel.ru) joined #forth
21:55:33 --- quit: chunkypuffs (Quit: ZNC 1.7.1 - https://znc.in)
21:56:11 --- join: chunkypuffs (~chunkypuf@2a01:4f9:2b:16d5::1) joined #forth
22:52:57 --- quit: proteusguy (Ping timeout: 250 seconds)
22:55:23 <dave0> nap time
22:55:42 --- quit: dave0 (Quit: dave's not here)
23:59:59 --- log: ended forth/19.04.28