00:00:00 --- log: started forth/21.05.26 00:03:15 --- quit: rber (Ping timeout: 245 seconds) 00:31:00 --- nick: xek_ -> xek 01:46:55 --- quit: Bogen85 (Read error: Connection reset by peer) 01:47:21 --- join: Bogen85 joined #forth 03:22:11 --- quit: cess11 (Quit: WeeChat 3.1) 03:24:52 hello, is there a Forth implementation that support POSIX threads ? 03:35:17 apparantly gforth can http://www.forth.org/svfig/kk/11-2018-Harbold.pdf 03:44:29 --- join: rber joined #forth 03:53:28 never dabbled in it but at the very least you should be able to reach out to some pthread c-library 03:56:47 ah ho! 03:56:51 ok 03:57:02 is the comp.lang.forth newsgroup active? 03:57:48 I come from #scheme, I made a prototype compiler targeting web assembly, but then figured, w3c is wasting a lot of effort put into forth. 03:58:10 well, they are different use cases, but but... 03:58:42 I mostly investigating a very long and broad project (window shopping basically) 03:58:50 I am not interested in the web stack. 03:59:26 --- join: neuro_sys joined #forth 03:59:48 so, I wondering whether forth and a particular implementation might not be better as target than relying directly un GNU lightning or LLVM/GCC JIT? 04:00:11 I mean rely on a forth as portable target. 04:01:33 I will ask my questions differently: A) Are their any language targeting forth, that are not forth themselfs? 04:02:08 B) outside gforth, what are other major forth implementation that works on netbsd and linux? 04:04:10 --- join: rtdos joined #forth 04:04:10 --- join: proteanthread joined #forth 04:05:49 the newsgroup is active, though heavily trolled :( 04:06:49 never heard of a transpiler to forth, seems like a weird detour to make some machine code 04:07:33 I wonder what would be the benefit? 04:07:39 (a) I've not run into much. I created a PL/0 backend that generates Forth some years ago, and there have been some efforts around transpiling C to Forth 04:08:30 someone else might correct me but in my experience non-trivial forth tends to be similar to programming with macros in lisps, one basically invents a programming language on the fly but looks at what it compiles to rather than stepping through expands 04:08:35 (b) how would one define major? 04:08:55 re major: wrong question. 04:09:08 i mostly use pforth, it's widely available 04:09:10 I mean, is there a forth that can produce linux / netbsd executables? 04:10:04 better question: what are performant forth? 04:10:45 one that isn't implemented in javascript? 04:11:21 i'm not sure what the use case is 04:12:26 some say gforth is slow, i've never felt it 04:14:05 neruo_sys, there would be an interest to rely on forth as a target, if there were forth implementations that can themself target various hardware architectures. Basically, in similar vein as LLVM IR except Forth hence stack-based. 04:14:34 other than gforth and the commercial ones, I'm not aware of anything that's focused on performance 04:14:53 and producing standalone executables is rare 04:15:23 mark4's x4 can do turnkey executables, but it's linux only 04:15:55 another question more related to stack machine. Why does all the CPU I know rely on registers? 04:16:27 --- quit: Zarutian_HTC (Ping timeout: 265 seconds) 04:16:37 is there a engineering / scientifical rationale for prefering register cpus? 04:19:07 Uneducated guess: I'd think registers come in handy for compiler optimizations (reusing several values across execution blocks). 04:19:41 Or it's just that the industry/market evolves in mysterious ways. 04:25:59 that's my impression too, don't have to push or dup data to use it, it can just lay idle in the register until outdated 04:25:59 I think the same 04:26:19 s/outdated/obsolete 04:27:13 but there are stack machine applications, some spacecraft runs in part on hardware forth interpreters 04:27:33 rtx2010 is the most well known i think 04:42:12 --- quit: cbridge-libera (Remote host closed the connection) 04:44:00 --- join: cbridge joined #forth 04:44:15 --- nick: cbridge -> Guest66358 04:46:37 --- quit: Guest66358 (Remote host closed the connection) 04:48:15 --- join: foucist joined #forth 04:50:03 --- join: cbridge-freenode joined #forth 04:52:01 --- quit: cbridge-freenode (Remote host closed the connection) 04:53:58 --- join: cbridge-freenode joined #forth 04:56:55 --- quit: neuro_sys (Ping timeout: 250 seconds) 05:00:48 I had to change the bridging bot name; the new management has been doing some bans/quiets on nicks which mention libera. 05:03:15 sure can do turnkey :) 05:03:23 but yat linux only :/ 05:03:37 but i suppose someone with a clue could port to windows probably 05:04:47 only reasin im awake this early is because roommate is flying to austin 05:04:54 --- quit: foucist (Quit: Changing server) 05:08:21 --- join: Zarutian_HTC joined #forth 05:10:18 test 05:10:50 hi foucist 05:11:11 I was testing teh bridge :P 05:11:31 I see you from the freenode side 05:11:35 we could tell :) 05:11:37 I think some of the windows forths produce executables 05:11:39 yea freenode sees you 05:11:53 win32forth does last i looked 05:12:23 ah I see, I accidentally disonnected from freenode while setting up liberachat 05:12:56 --- join: Zarutian_HTC1 joined #forth 05:12:56 --- quit: Zarutian_HTC (Read error: Connection reset by peer) 05:12:57 re: registers vs stack machines - you really bottleneck your performance confining things to a stack since you can only do operations on the top of it 05:13:06 this may be interesting: https://en.wikipedia.org/wiki/Stack_machine#Performance_disadvantages_of_stack_machines 05:13:12 not in my experience 05:13:41 i think that one thing does slow you down but other things about how forth executes compared to say how c executes make up for that 05:13:54 I dont think that is the case at all 05:14:02 the fact that forth does not bury its parameters under the return address etc... i think it balances out :) 05:14:20 at least ive never seen those numbers. its also difficult to imagine how that could ever be true when you start thinking about the assembly that actually gets executed 05:14:57 yea what REALLY needs to be done is two people need to get toteher and hash out a design for something non trivial 05:15:02 mark4, the parameters stay in the registers on a lot of architectures and only get put on the stack as necessary which may not be much or at all 05:15:11 they then both write it in forth and then both write it in c colab not indivually 05:15:17 and then the results of both can be compared 05:15:34 ya, I talked to tp about that in #mecrisp and he was interested in that 05:15:53 on aarch64 a significant number of opcodes are move reg, reg and it looks horrendous 05:16:04 who cares what it looks like 05:16:10 both people whould have to work on both versions :) 05:16:11 youre not writing it 05:16:12 --- quit: Zarutian_HTC1 (Read error: Connection reset by peer) 05:16:17 --- join: foucist joined #forth 05:16:25 it looks horrendously inefficient lol 05:16:26 LOOKS 05:16:31 but maybe not 05:16:35 ya definitely not 05:16:46 literally 8 mov reg, reg then a call foo then 8 mov reg, reg 05:16:49 all over the place 05:16:57 --- quit: cbridge-freenode (Remote host closed the connection) 05:17:15 I would love to see an example of that with optimizations enabled 05:17:59 just out of curiosity 05:18:04 --- join: cbridge-freenode joined #forth 05:20:26 its been a while since i ran a aarch64 compile and then it was android ndk 05:20:49 and its also possible that the repeated reg, reg moves are no longer needed 05:21:08 but i did observe them and they horrified me lol 05:26:35 --- join: neuro_sys joined #forth 05:26:59 --- nick: neuro_sys -> Guest76084 05:27:07 --- nick: Guest76084 -> neuro_sys 05:27:13 --- quit: neuro_sys (Changing host) 05:27:13 --- join: neuro_sys joined #forth 05:34:02 I'm fancying an emacs-like written in Forth... 05:34:14 editor* 05:34:47 Emacs is a bit too slow, and feels already a bit bloated 05:35:53 when a game says xyz frames per second does that mean the entire scene pixle for pixle is redrawn every scene or does the video card only update the actual chaned pixles? 05:36:03 I'm going to be working on my editor soon. 05:36:22 I have a decent starting point I wrote last go round; source lives in one 4k block. 05:36:30 But I have more plans for the next iteration. 05:36:52 It does indeed use emacs-like keystroke mechanics, but with vim key assignments. 05:37:10 I.e., ctrl- is how all of the controlling gets done. 05:37:15 mark4: Usually all pixels change considering most realtime graphics involve constant camera movement. 05:38:12 right but im saying my text mode ray casting engine gets 20k frames per second on my real gaming machine - and im only updating the changes. is that cheating? 05:38:16 and thats WITH movement 05:38:23 maybe uEmacs / microemacs could be a source of inspiration? 05:38:36 It's not cheating at all, it was the standard back in the day. 05:38:43 The GPU may be doing some similar optimizations if possible. 05:39:35 i am redrawing every single pixle (1 bit) intoa frame buffer which i then translate into utf8 braile characters which i then write to a window in my TUI, that window is then drawn into a screen FULLY, only the differences between the new screen and the old screen are written to the console 05:39:42 is calling that 20k fps cheating? 05:40:33 I don't think it is. Just to clarify, in the worst case, as in when the scene has entirely changed, all of the pixels will change right? 05:41:05 20K is great but okay for a graphical window system I guess. 05:41:10 well its a ray casting engine so im drawing one vertical stripe down every colum of the scene for based on wall height 05:41:18 Ah a ray casting engine 05:41:24 i have to driave roomie to the airport 05:49:36 --- quit: foucist (Read error: Connection reset by peer) 05:50:07 I'm confused - whose eyes can utilize 20k frames per second? 05:50:31 I sure don't want to meet THAT in a dark alley. 05:51:02 --- quit: cbridge-freenode (Remote host closed the connection) 05:51:11 Or do you mean you can generate a video at a 20k per second rate? 05:51:24 The monitor can't display that many frames anyway :) 05:51:32 Right. 05:51:50 So we must not be talking about realtime rendering. 05:52:00 The CPU will be crunching scenes for no reason most of the time, and there'll be some tearing artifacts in the middle of the frame occasionally. 05:53:08 Speaking of realtime graphics and Forth: https://www.youtube.com/watch?v=h829mkWXpYM 05:53:21 --- join: cbridge-freenode joined #forth 05:54:09 --- join: f_ joined #forth 05:57:55 high fps may be wasteful in practice, but it can be a metric used to measure the draw speed 05:58:07 Sure. 05:58:42 I just wandered in in the middle of the conversation. :-) 05:59:02 and i wandered in at the end, so 06:18:59 --- quit: f_ (Read error: Connection reset by peer) 06:22:44 --- join: f joined #forth 06:23:07 --- nick: f -> Guest87537 06:29:24 --- part: guan left #forth 06:48:48 --- quit: Gromboli (Quit: Leaving...) 06:58:42 --- join: Zarutian_HTC joined #forth 07:05:26 -!- f-a(~f-a@151.82.110.98) has left ##forth 07:09:45 echo 07:10:00 well this works 07:25:13 --- part: rann left #forth 07:35:42 amirouche: I'm working on a compiler to Forth, but it's a bit more meet-in-the-middle than that; I'll be compiling an ANFified version of my language to a bytecode derived for it, then implementing that bytecode "natively" in Forth (i.e. as Forth words, rather than writing a separate intepreter in Forth) 07:38:00 I think another strong reason for register machines is it's *way* easier to do out-of-order execution and other performance tricks with them 08:03:06 Oh dear you've been taken in by this freenode drama then? 08:04:10 I am so tired of bridges from discord poaching, now I have to put up with surly mods splitting the network 08:05:56 veltas, won't surprise me if libera dies after not too long and then the bridge can go. Right now I just care about keeping our forth community together. Freenode #forth has really been doing well the last couple of years and it's horrible to see this nonsense damage that. 08:06:12 Absolutely 08:06:49 Freenode isn't about the mods, it's just the network that happened to get a lot of people. As long as mods aren't shit and don't k-line me accidentally too much we're golden 08:07:15 freenode are literally taking over people's channels 08:07:44 I don't understand why you stick around on a network that is banning people for talking about migration, deoping users in their own channels (including #ubuntu and #haskell for example) etc 08:07:48 If you were on freenode you'd know that was a mistake and reaction to the mods trying to ruin the network 08:07:52 freenode is a dumpster fire 08:08:05 is taking over 700+ channels a mistake? lol 08:08:07 Well really you've got to read between the lines 08:08:17 channels that just so happen to be ones where people are discussing migration to another network? 08:08:35 because freenode might *just so happen* want to keep new users oblivious to the drama to stop them from leaving? 08:08:38 by taking over channels? 08:08:45 Well I wouldn't advise you even read their explanation because you've already made your mind up 08:08:55 I have read freenode's explanation 08:08:58 the one on the site? 08:09:00 it's half-assed BS 08:09:04 I happened to read both sides and realised I care a lot less about either side than trying to keep things in one place 08:09:27 It's not about who's right or wrong, it's about where my communities are 08:12:42 But I believe it was a mistake, they were reacting to mods that were banning everyone on freenode they could to try and force a move 08:12:42 There are indeed two sides but it requires not just agreeing with whatever people on hacker news or twitter say 08:12:43 None of this would have happened if self-important mods hadn't tried to do a split, over stuff that really doesn't seem to be that important compared to the communities. The perfect always the enemy of the good 08:12:43 btw, reading n.y.c is not good for ones health (And no I do not consider that site to have "hacker news") 08:12:43 nihilazo, while freenode's tactics may not be the best - they are defending themselves against an existential attack on their existence so I'm cutting them a bit of slack so long as they don't break the channels I'm on. And would give them a chance to repair should they do so. 08:12:43 Exactly 08:13:01 I don't think Freenode team is "in the right", they certainly share blame for the drama, but it's just drama at the end of the day 08:13:01 --- quit: Zarutian_HTC (Read error: Connection reset by peer) 08:13:13 --- join: Zarutian_HTC joined #forth 08:13:51 yep - no one has handled this well. But someone clearly threw the first grenade at my community and so they will not be forgiven. Time will only tell if freenode will keep it together. 08:15:11 it was andrew lee who make the first offensive, the mods reacted 08:15:27 and freenode has been a sinking ship ever since 08:15:43 even if their moves make sense in terms of protecting their self-interest, they do not make sense as a good place for communities 08:15:49 you don't oust opers from their own channels 08:15:59 or take over people's communities 08:16:10 unless, of course, you are freenode 08:16:24 (and I have looked at what freenode have to say for themselves. I just disagree) 08:16:41 "sinking ship" You know as well as I do it will be just as active in a year 08:16:51 nah 08:16:53 it won't 08:17:01 the big communities have all left 08:17:07 or are all in the process of leaving 08:17:14 nihilazo, sorry but andrew didn't do anything to #forth. He may have made policies to cause his employees to leave but that's between them. When it splashed on my channel it was libera that caused that. They could have consulted the communities before starting destruction but their sense of entitlement showed they didn't give a damn about you or me. 08:17:23 freenode will stay active but the hub of open source it will no longer be 08:17:25 Exactly 08:17:30 Self-important mods 08:17:41 Fuck Andrew Lee but he didn't pull the pin 08:17:46 truly 08:18:53 Freenode is going to keep being active, and it's not really ever been the hub of "open source" 08:18:55 BTW - I've made a #forth-rant channel. I suggest if people wanna discuss the irc node nonsense we do it there. Let's keep #forth on things that at least involve stacks. 08:19:18 Sorry I'll stop now 08:19:26 veltas, I'm just as guilty... 08:21:07 --- quit: neuro_sys (Remote host closed the connection) 08:22:09 so #forth on freenode is banned? 08:22:13 reminds me of a MUD I knew of, it was very very hard to get wizard (read: op) privs there. And anyone who sought them was disqualified from getting any for a long while. It cut down on such drama quite a bit. 08:22:56 ldb nope- it's working perfectly. 08:24:47 --- join: ldbeth joined #forth 08:25:37 ah, good then. 08:31:43 what are the popular forth implementations right now? last time I checked there're GNU's Forth and 4th 08:33:30 remexre: what forth do you use? 08:33:41 remexre: what forth implementation do you use? 08:34:51 amirouche: wrote my own :) https://git.sr.ht/~remexre/stahl 08:35:23 er, https://git.sr.ht/~remexre/stahl/tree/main/item/kernel/src/common is the more relevant subdir 08:35:25 "writing your own" is by far the most popular implementation! ;-) 08:37:50 I see that is a charming point 08:39:05 ldbeth, problematic when you're trying to do stuff with it, however. I'm starting my 3rd refactor of actorforth... haha but it is promising and fun. Clearly destined to take over the world no doubt. ;-) 08:39:35 remexre: reading on stalhdream 08:40:00 -!- ldb(~ldbeth@user/ldb) has left ##forth 08:47:30 remexre: I hit ENTER too quickly, where can I follow you progress, do you have sourcehut ml? 08:48:15 recently I read Algorithms for Functional Programming and the Scheme code inside heavily uses values and call-with-values which reminds me the forth 08:53:08 --- part: Guest87537 left #forth 08:54:45 ldbeth, it's amazing how being pointfree gets rid of a lot of the cruft of dealing with variable names in functional languages. Our stack juggling is their name juggling. And their lists are almost always really used as stacks! ;-) 08:56:47 remexre: I have a similar project using scheme, and okvs for the database. 09:00:48 proteusguy: true, no point to get the type 09:01:33 ldbeth, in actorforth the type is attacked to the stack. It's a strongly typed language. 09:02:33 attached... not attacked. :-) 09:03:17 --- quit: ldbeth (Quit: leaving) 09:03:51 --- join: ldbeth joined #forth 09:04:29 proteusguy: that makes sense, it is painful handling strings in forth 09:05:07 it can be indeed. 09:20:51 Yes. I want to make improvements on that in mine - it's one of my big goals. 09:21:13 And I want a streamlined regular expression capability too. 09:21:52 I like how Python handles strings. 09:22:10 But that can't just be "bolted on" to Forth. 09:22:21 The "integration" needs attention. 09:27:49 --- quit: cbridge-freenode (Remote host closed the connection) 09:28:21 --- join: cbridge-freenode joined #forth 09:43:29 ok :) next milestone is probably either the functional lang getting finished (the backend is basically done, the frontend keeps giving me trouble...) or the OS part getting 09:43:52 a proper scheduler 09:44:06 siraben: yeah lol didn't get too far 09:48:05 remexre: ah, what was the goal? 09:48:08 --- quit: Zarutian_HTC (Ping timeout: 240 seconds) 09:48:09 to verify the assembler? 09:50:40 https://img.vim-cn.com/2e/ecebc31762419f7d2684522c2633400048881b.png 09:50:48 my very cheap forth 09:55:39 ldbeth: huh, looks typed? 09:58:16 very cool 09:58:20 siraben: sort isn't type, it applies to syntax only 09:58:30 ah, I see 10:13:38 siraben: yeah; but nothing being inductive anymore means the tactics I'm used to don't work, so I sorta gave up until I learn Ltac 10:13:52 ldbeth: ooh, I haven't seen maude before; neat! 10:15:00 remexre: I'm dipping my toes into Ltac, it's quite interesting when you can pattern-match goals and do tactics 10:22:39 --- join: ldbeth` joined #forth 10:25:35 --- quit: sts-q (Ping timeout: 260 seconds) 10:25:36 --- quit: ldbeth (Ping timeout: 260 seconds) 10:26:07 --- quit: Guest78830 (Ping timeout: 260 seconds) 10:26:52 --- join: WilhelmVonWeiner joined #forth 10:27:00 --- join: sts-q joined #forth 10:27:16 --- nick: WilhelmVonWeiner -> Guest21384 10:27:28 I read that and first tried to parse it as a discussion of electronic circuitry (... inductive...). 10:27:31 But... no. 10:31:08 --- quit: ldbeth` (Quit: ERC (IRC client for Emacs 27.2)) 10:36:02 --- join: Zarutian_HTC joined #forth 11:32:33 --- quit: boru (Quit: Good riddance.) 12:23:45 so for a naming question: retro has some words that were borrowed from factor. These are bi bi* bi@ tri tri* tri@ 12:23:45 I was wondering if anyone has suggestions for better names for these 12:23:45 http://forth.works/share/032e1a7fe154073270b003de9e42d796 is the summary documentation for them 12:25:45 --- quit: cbridge-freenode (Remote host closed the connection) 12:45:22 lol my text mode ray casting engine now does texture mapping for the walls :P~ 12:45:24 floor next 12:48:24 --- join: cbridge-freenode joined #forth 12:51:45 crc: English doesn't have good words for this, but the idea reminds me sliiightly of SIMD (or the converse, which I guess would be MISD). Maybe there is some jargon from that field that you could steal? 12:52:54 Otherwise, using your own made-up, short words, (like you are doing now) is probably the best option. 12:57:29 Now, the next question is how to distinguish between the three flavors of "bi" and "tri", which are doing quite different things. I think of them as: (1) copied misd (bi), (2) in-place misd (bi*), (3) simd (bi@). 12:59:33 The mnenomics that make sense to my brain are: 2misd, 2micd, 2simd, 3misd, 3micd, 3simd. Where misd is "multiple instructions, single data" and micd is "multiple instructions, copied data", and the leading number is the multiplicity. 13:00:22 But, I have no idea if anyone else's brain works that way. And I have also never used Retro so I don't know how well this fits into your vocabulary. 13:03:51 Wait, I misunderstood how bi* and tri* work. They are not misd, instead they apply the N given functions against the N given parameters. Hm.... This means the above is probably not good advice. Sorry about that. 13:23:06 --- quit: mark4 (Ping timeout: 265 seconds) 13:29:27 psf: no worries; I've not looked at SIMD before, so this is interesting :) 13:29:59 --- join: mark4 joined #forth 13:39:14 Maybe there is one salvageable bit of advice in here. The Factor names are two-part, containing both a multiplicity (bi, tri) and an operation (blank, *, @). The operations are opaque symbols which must be memorized. 13:39:56 I think this would work better the other way around. Use a numeric prefix (2 or 3) for the multiplicity, and come up with short descriptive names for each operation. 13:40:27 To my eyes, this is more Forthy. Consider for example dup vs. 2dup. 13:41:57 --- quit: gravicappa (Ping timeout: 265 seconds) 13:42:17 --- join: gravicappa joined #forth 13:43:03 It assigns more characters to the "hard to understand" part of the name, so might be more memorable. Unless those sigils * and @ are used consistently in other parts of the language so one can count on the programmer to have already memorized them. 13:46:58 --- quit: gravicappa (Ping timeout: 260 seconds) 13:46:58 --- quit: Zarutian_HTC (Read error: Connection reset by peer) 13:47:02 --- join: Zarutian_HTC1 joined #forth 13:47:24 --- join: gravicappa joined #forth 14:01:06 --- quit: xek (Read error: Connection reset by peer) 14:01:06 --- quit: Bogen85 (Remote host closed the connection) 14:01:22 --- join: Bogen85 joined #forth 14:01:35 --- join: xek joined #forth 14:02:38 This sounds interesting - I missed what preceeded the bridge joining. 14:08:12 KipIngram: I'll repost it for you. 14:08:14 12:23 <@crc> I was wondering if anyone has suggestions for better names for these 14:08:19 12:23 <@crc> http://forth.works/share/032e1a7fe154073270b003de9e42d796 is the summary documentation for them 14:08:31 Taking a look... 14:09:41 What does the string of letters after Data: tell me? 14:10:45 Not a retroforth expert, but to my eyes, it looks like a condensed stack comment with one letter per parameter. 14:11:31 Yeah, that's how I'm trying to interpret it. I haven't ever delved into retro either. 14:11:49 Factor calls these cleave, spread, and apply combinators 14:11:56 Not sure what the -? is representing. 14:11:59 https://docs.factorcode.org/content/article-cleave-combinators.html 14:12:09 https://docs.factorcode.org/content/article-spread-combinators.html 14:12:16 https://docs.factorcode.org/content/article-apply-combinators.html 14:12:44 -? Means the return values are unknown 14:12:49 Q is a pointer 14:13:14 Others are general values on the stack 14:35:09 Thanks for the explanation! 14:35:49 --- quit: proteus-guy (Ping timeout: 272 seconds) 14:36:37 Those names "cleave", "spread" and "apply" are not bad, if a bit long. Also, "apply" is a lot of "map" in functional languages. 14:37:47 long names aren't an issue for me 14:39:01 My average word name is 7 characters; with the longest being 25-30 14:39:49 Then maybe 2cleave, 2spread, 2apply and 3cleave, 3spread, 3apply? 14:40:35 Though I'm not the wildest about those names. 14:41:02 I'll try that and see how it feels to use in actual code 14:41:38 Naming is hard sometimes :( 14:42:28 I haven't checked my name statistics, but I'm pretty sure it's less than that. 14:42:35 By a fair bit. 14:48:29 --- join: proteus-guy joined #forth 14:58:27 some of that is name spacing related (s: for strings, file: for files, etc), but my names are probably longer than most forths 15:02:39 My personal Forth does the Forth Inc. trick of storing just the first 3 characters plus the length. 15:03:02 that definitely would not work in my system 15:03:02 But if I wrote sizeable programs, I would... not be able to get away with that. :P 15:03:26 Yeah, exactly. Diff'rent strokes. 15:03:49 My largest applications are a few thousand words; way too much overlap, even with more concise naming :) 15:04:20 It's a good fit for a spartan 64k DOSforth, but not good for a large capable system with many builtins! 15:04:28 (Like yours) 15:07:47 It's a nice thing about Forth. It scales well from minimal systems to big ones. 15:10:22 Indeed. 15:14:11 on Forth I miss some decent one for termux (android) 15:15:07 I tried pforth but it crashed on some stack operations 15:15:16 such as >r swap r> 15:19:14