2024-09-18 00:04:14 <zelgomer    > i recently adopted something i saw from muforth. my ] includes an implicit "literal". you have to use -] to start compiling again without compiling a literal. just wanted to share, this is a pretty major quality of life improvement.
2024-09-18 00:05:09 <amby        > i have ], for that
2024-09-18 00:05:58 <zelgomer    > yeah that makes sense too
2024-09-18 00:25:09 <xentrac     > seems reasonable
2024-09-18 00:44:04 <zelgomer    > i think i have a good idea about how to do what i want to do, and as usual, i'm finding it difficult to proceed because i'm terrible at naming things
2024-09-18 00:47:00 <amby        > two hard problems etc etc
2024-09-18 01:11:03 <zelgomer    > how many words can i write named <some-single-letter-prefix>@ ?
2024-09-18 01:11:23 <xentrac     > it depends on which alphabet you're using and whether you're case-sensitive
2024-09-18 01:11:31 <zelgomer    > lol
2024-09-18 01:11:41 <xentrac     > ø@ and ÿ@ for example may not be easy for you to type
2024-09-18 01:12:05 <zelgomer    > i use vim, so digraphs are not too far away
2024-09-18 01:12:14 <xentrac     > and λ@ and Ω@ may be even worse
2024-09-18 01:12:23 <xentrac     > not to mention ℵ@
2024-09-18 01:13:57 <zelgomer    > : ^kl*@ ...
2024-09-18 01:15:51 <amby        > define letter
2024-09-18 01:16:19 <amby        > printable ascii character? unicode codepoint? grapheme cluster?
2024-09-18 01:54:54 <zelgomer    > well tbh my forth only understands ascii
2024-09-18 01:58:52 <xentrac     > can it tell the difference between ASCII and UTF-8, though?
2024-09-18 02:02:57 <amby        > unless your forth does its own text rendering, it probably can't
2024-09-18 02:03:22 <amby        > well. CHAR and EMIT are slightly sticky
2024-09-18 02:08:10 <xentrac     > that's a good point.  Also I've had problems with UTF-8 messing up the width of 64-byte block lines
2024-09-18 02:09:22 <amby        > you might end up with CCHAR and CEMIT for when you specifically wanna deal with bytes
2024-09-18 02:09:30 <thrig       > unicode can also have some fun normalization problems
2024-09-18 02:09:54 <amby        > and also maybe CHAR, for compiling a unicode codepoint as utf-8
2024-09-18 02:11:04 <xentrac     > I think Forth 2012 has some stuff about this
2024-09-18 02:11:31 <xentrac     > https://forth-standard.org/standard/xchar
2024-09-18 02:11:50 <xentrac     > not that this is the only possible way to handle the problem, but it's one way
2024-09-18 11:58:26 <veltas      > CHAR works fine for ASCII and S" works for UTF-8, EMIT should work with UTF-8 just fine, just not if you're trying to count columns or character indices
2024-09-18 11:58:46 <veltas      > S\" should also work with UTF-8
2024-09-18 11:59:59 <veltas      > Or any parsing really, CHAR and [CHAR] are the only parsing words I know that take the first byte
2024-09-18 12:00:15 <veltas      > And it's not much of a limitation that PARSE can only use ASCII as a delimiter
2024-09-18 12:39:05 <xentrac     > well, decimal 9786 emit is unlikely to produce a happy face unless emit is internally doing the UTF-8 encoding
2024-09-18 12:41:37 <xentrac     > but you could decimal 226 emit 152 emit 186 emit if that's what you mean
2024-09-18 13:10:51 <veltas      > You can do that but you can also do S" <insert horrifying emoji>" and TYPE or loop-EMIT it out and it will work
2024-09-18 13:11:00 <veltas      > If you want to get a codepoint from UTF-8 encoding that's not hard, but it's also totally unnecessary or even misguided for most work
2024-09-18 13:34:26 <xentrac     > it can be useful sometimes
2024-09-18 13:34:53 <xentrac     > but it definitely has its pitfalls
2024-09-18 13:39:34 <veltas      > How to approach localisation in Forth:  1. Don't.  2. Dön't.
2024-09-18 13:40:14 <olle        > xD
2024-09-18 14:03:46 <amby        > do, but only if you need to
2024-09-18 14:08:11 <olle        > Easier to just teach English to the entire world.
2024-09-18 14:08:43 <veltas      > What O'Reilly animal is appropriate?  "MVC with Forth"
2024-09-18 14:09:51 <veltas      > "Agile Front End Web Development with Forth"
2024-09-18 14:11:17 <olle        > Something extinct? A mammoth?
2024-09-18 14:11:33 <olle        > Or a bear animal, what's it called in english
2024-09-18 14:11:43 <olle        > Hardened to live under extreme pressure in outer space or vulcano
2024-09-18 14:11:51 <amby        > tardigrade
2024-09-18 14:11:54 <olle        > Right
2024-09-18 14:12:05 <amby        > cockroach probably gets the same message across clearer
2024-09-18 14:12:09 <amby        > and is slightly more photogenic
2024-09-18 14:12:42 <olle        > ^^
2024-09-18 14:13:10 <olle        > https://www.1-9-9-1.com/
2024-09-18 14:21:35 <veltas      > Extinct? No, mythical
2024-09-18 14:21:42 <veltas      > Like a chimera
2024-09-18 14:22:46 <olle        > Unicorn?
2024-09-18 14:22:56 <olle        > MVC is ded anyway
2024-09-18 14:22:59 <olle        > It's all pipelines now
2024-09-18 14:23:10 <veltas      > "Game Engines with Forth in Rust Macros"
2024-09-18 14:23:53 <veltas      > "SIMD Optimisation in F83"
2024-09-18 14:29:19 <veltas      > "SwiftForth Microservices"
2024-09-18 14:31:17 <olle        > Well
2024-09-18 14:32:12 <veltas      > I've applied MVC on the ZX Spectrum, is that cursed?
2024-09-18 14:32:34 <user51      > "Busy Beaver solved using a stack machine"
2024-09-18 14:32:58 <olle        > :d
2024-09-18 14:33:50 <veltas      > olle where can I read about pipelines?
2024-09-18 14:34:14 <veltas      > I don't know a thing about services or the web
2024-09-18 14:37:47 <olle        > veltas: perhaps this? https://doeken.org/blog/middleware-pattern-in-php
2024-09-18 14:38:02 <olle        > Middleware is more onion-like that pipelines. Still.
2024-09-18 14:38:50 <xentrac     > yeah, tardigrades are clearly the right theme animal for Forth
2024-09-18 14:39:23 <olle        > A uni-tard
2024-09-18 14:39:30 <olle        > Uni-turd lol
2024-09-18 14:39:33 <olle        > Nah
2024-09-18 14:39:41 <xentrac     > Unicorns are Unicon
2024-09-18 14:42:26 <xentrac     > add GC and static types and I think you have maybe Cat, or Kitten?
2024-09-18 14:44:56 <veltas      > This concept you linked to me olle feels like using vectorised exection on TYPE and KEY
2024-09-18 14:45:00 <veltas      > execution*
2024-09-18 14:45:29 <veltas      > Certainly that's how I use it, often I just extend the existing execution tokens in there
2024-09-18 14:45:36 <veltas      > with post actions and pre actions
2024-09-18 14:46:31 <veltas      > Double Pass sounds like LaTeX and two-pass assemblers
2024-09-18 14:47:19 <veltas      > Interesting anyway
2024-09-18 14:47:33 <veltas      > I do feel like we're all resolving the same problems in different levels of the software stack
2024-09-18 14:51:12 <xentrac     > with more or less unnecessary complexity and observability
2024-09-18 14:51:38 <olle        > There's only so many design patterns you need to solve 99% of the problems.
2024-09-18 14:52:09 <veltas      > re-solving*, not resolving
2024-09-18 20:29:03 <user51      > 
2024-09-18 20:31:55 <user51      > I've been trying to do neat stuff but it's only neat in theory  :/
2024-09-18 20:34:29 <user51      > Yeah, I knew I'll get get an idea once I type it on IRC.  That's some freaky XOR logic there.
2024-09-18 20:38:15 <thrig       > a rubber ducky can also work
2024-09-18 20:39:04 <user51      > I tried so hard and in the end..  :)
2024-09-18 20:39:31 <user51      > I was trying to figure it out while sketching on my ipad, didn't work.  Guess IRC is still the best for that.
2024-09-18 20:39:59 <thrig       > or take a walk or something
2024-09-18 20:41:48 <user51      > Did that earlier, too.  I'll stop here before someone thinks, man, this guy only has one stack.
2024-09-18 20:52:31 <xentrac     > haha
2024-09-18 22:37:07 <KipIngram   > user51: Nice.  :-)
2024-09-18 22:37:13 <KipIngram   > I'm gonna remember that one.