2023-10-22 03:13:56 <veltas      > KipIngram: You could probably write something similar to a FORTRAN 2 compiler yourself, or a B compiler, or lisp interpreter
2023-10-22 03:14:13 <veltas      > In assembly I'd wager
2023-10-22 03:14:29 <veltas      > Old COBOL too
2023-10-22 03:15:01 <veltas      > Forth is interesting in the regard that it never got *more* complicated to write
2023-10-22 03:15:04 <GeDaMo      > https://justine.lol/sectorlisp2/
2023-10-22 04:03:09 <rendar      > what you think folks about Factor?
2023-10-22 09:13:02 <zelgomer    > rendar: i appreciate the spirit, don't like the execution
2023-10-22 09:14:02 <zelgomer    > all this listener stuff and scaffold... i just want to open a text editor of my choice and start typing
2023-10-22 11:40:28 <KipIngram   > I mostly agree, zelgomer, though I do think a Forth environment offers some nice opportunities for things like auto-complete and so on.
2023-10-22 11:40:43 <KipIngram   > Not "requisites," but rather just "perks."
2023-10-22 11:41:26 <KipIngram   > Things you'd gain by having the editor a little more "integrated" into the environment.
2023-10-22 11:45:56 <KipIngram   > The way you phrased that implied a fully "offline" source preparation process - I do usually think of my editor running *within* the Forth system as part of it.  Forth being interactive in the way that it is and all.
2023-10-22 11:47:53 <KipIngram   > One of Chuck's old systems had a highly integrated editor - as he typed things the editor would search them, and if the word was a "standard built in" word those source chars would get replaced by a "token" right there in the source.  He was heavily focused at the time on minimizing the size of his source.  I don't think that is quite as important these days as it may have been to him at the time.
2023-10-22 11:48:46 <KipIngram   > But it did make the source take up less space and also made it much faster to compile - he later didn't need to search for words that had gotten that treatment at all; that token gave him easy access to the word's CFA.
2023-10-22 11:49:36 <KipIngram   > He was "almost" able to just copy straight from source to dictionary.
2023-10-22 11:51:15 <rendar      > zelgomer, can you elaborate it? why you don't like Factor execution?
2023-10-22 12:02:06 <KipIngram   > veltas: Yeah, I probably could do those things, but they mostly aren't interactive in nature.
2023-10-22 12:03:50 <KipIngram   > I guess that's another thing I love about Forth - how good it is for poking at new hardware you're working with.
2023-10-22 12:04:14 <KipIngram   > I've always felt like embedded work is Forth's real "natural home."
2023-10-22 13:06:36 <zelgomer    > rendar: i feel like i did elaborate. which part of my second message did you need me to explain further?
2023-10-22 13:09:02 <zelgomer    > note i'm really a c guy. for me to adopt an entirely new ecosystem, unless it's something that i can write myself from scratch, the bar is pretty high there. i don't know why i would just use c, since it's already a complicated ecosystem written by other people, but at least i already understand it
2023-10-22 13:09:15 <zelgomer    > what appeals to me about forth is the dream of rolling my own
2023-10-22 13:09:26 <zelgomer    > factor doesn't give me that
2023-10-22 13:10:24 <zelgomer    > i'm not a "gadgets" guy. i'm not impressed by fancy guis and instruments and things. i'm a minimalist, and the factor development environment doesn't feel very minimalist to me
2023-10-22 13:11:00 <rendar      > zelgomer, oh sorry i missed that
2023-10-22 13:15:40 <rendar      > zelgomer, now i perfectly got your thoughts and i absolutely agree
2023-10-22 13:18:42 <KipIngram   > zelomer: You can definitely write a Forth.  You can do it in C, and if you use gcc then you can make it "tick" under the hood in a fully "legitimate Forth way."  I wrote one that way once.  The primitives were C "one-liners," but otherwise it was a full-on properly operating Forth.  These days I prefer to use assembly, but either way works fine.
2023-10-22 13:19:03 <KipIngram   > gcc has that "pointer to label" extension that lets you write a proper inner interpreter.
2023-10-22 13:20:29 <KipIngram   > My C Forth wound up getting awfully ugly, source-wise, before I was done, though.
2023-10-22 13:20:57 <KipIngram   > I wound up embarrassed to show it to anyone, the source was so hideious.
2023-10-22 13:21:24 <KipIngram   > s/hideious/hideous/
2023-10-22 13:23:08 <KipIngram   > Oh, you can't necessarily get full performance with a C implementation, since you can totally control what you do with the registers.  That's the biggest argument for moving to assembly.
2023-10-22 13:23:23 <KipIngram   > "can't" totally control
2023-10-22 13:25:07 <rendar      > my actual Forth is written in Rust and it has exactly 473 builtin words, for now
2023-10-22 13:26:18 <KipIngram   > Rust is one I've obtained no knowledge whatsoever so far.  I just have gotten the impression it's popular in web-related applications, but even that might be a misperception.
2023-10-22 13:26:40 <rendar      > KipIngram, you're right, but Rust is growing a lot also in gaming, embedded, crypto..
2023-10-22 13:26:57 <rendar      > and webassembly
2023-10-22 13:27:10 <KipIngram   > Cool.  It seems to pop up a lot as a "dependency" in various packages I've looked at.
2023-10-22 13:27:18 <rendar      > it's worth learning it, it's the most loved lang
2023-10-22 13:27:57 <rendar      > KipIngram, one key feature of Rust is its ecosystem of packages, you'd have also very tiny packages to do small things
2023-10-22 13:31:59 <thrig       > and supply chain vulns! and horrifyingly slow compile times!
2023-10-22 13:34:52 <rendar      > thrig, there are not perfect things
2023-10-22 13:35:12 <rendar      > thrig, slow copmile times is true, but its solvable, supply chain vulns has been debunked
2023-10-22 13:40:41 <thrig       > what changed since, uh, August 28th when the last (reported, known) malicious packages were taken down?
2023-10-22 13:42:18 <KipIngram   > Well, I just looked up "rust supply chain vuln" and it looks like the idea is to get someone to download an "imitation" of some legit package?  I would hardly blame that on the language itself...
2023-10-22 13:43:21 <thrig       > sure, but if you've got something with a bazillion crates, then you have to go through each crate by hand to look for malicious code
2023-10-22 13:43:28 <rendar      > KipIngram, indeed
2023-10-22 13:43:33 <KipIngram   > Especially if a package name typo is required for it to work - users do need know what they're doing.
2023-10-22 13:43:38 <rendar      > also node.js has/had that plague
2023-10-22 13:43:50 <rendar      > people still uses node.js successfully
2023-10-22 13:44:00 <thrig       > and then there's "binary blob in serde_derive procedural macro crate" which seems hard to audit?
2023-10-22 13:44:27 <thrig       > also the rust portability is shit. maybe some decade it will nudge itself up to the "maybe usable" category?
2023-10-22 13:45:05 <rendar      > thrig, huh? I successfully compile Rust in AMD64, x86, ARM and webassembly
2023-10-22 13:45:23 <KipIngram   > :-)  I'm starting to get the impression that this is yet another topic in the tech world that precipitates holy wars.
2023-10-22 13:45:25 <rendar      > a lot of embedded products uses Rust to compile it in very esotic arch
2023-10-22 13:45:36 <rendar      > ahaha yeah KipIngram
2023-10-22 13:46:30 <thrig       > very high odds of a rust compile shitting itself after wasting who knows how much CPU on OpenBSD
2023-10-22 13:46:51 <rendar      > that never happened to me
2023-10-22 13:46:59 <rendar      > in 1.8 yrs i use Rust
2023-10-22 13:47:32 <thrig       > lucky for you
2023-10-22 13:55:53 <ghoulpine   > the serde_derive blob was removed (only after a bunch of backlash though, heh)
2023-10-22 13:58:00 <MrMobius    > holy war is right :P
2023-10-22 13:58:12 <MrMobius    > also infatuation when it comes to Rust but then again there is some of that here in Forth
2023-10-22 13:58:16 <ghoulpine   > thrig: i assume that probably means it ran out of ram? apparently the rust compiler itself is impossible to compile on 32 bit openbsd because it uses way too much ram >.>
2023-10-22 14:01:20 <KipIngram   > Forth *definitely* can ignite holy wars.  We just mostly all like it here.  Quite a few years ago there was a Russian guy that hung out here, though, nickname Asau, who seemed to take very chance to tear Forth down.
2023-10-22 14:01:27 <KipIngram   > I never understood why he hung out here.
2023-10-22 14:01:53 <KipIngram   > At first I'd try to debate him, but I eventually decided he was just trolling us.
2023-10-22 14:02:31 <KipIngram   > I think I recognize trolls more quickly these days - I was pretty inexperienced online back then.
2023-10-22 14:04:55 <rendar      > lol
2023-10-22 14:05:01 <rendar      > yeah there are many trolls
2023-10-22 14:05:11 <rendar      > and the main rule is: don't feed the troll :)
2023-10-22 14:12:20 <zelgomer    > KipIngram> zelomer: You can definitely write a Forth.
2023-10-22 14:12:23 <zelgomer    > i know. i have
2023-10-22 14:12:39 <zelgomer    > the question was about factor, not forth
2023-10-22 14:13:26 <rendar      > zelgomer, you wrote that in C?
2023-10-22 14:13:48 <zelgomer    > no
2023-10-22 14:13:59 <rendar      > in which language, if i may ask?
2023-10-22 14:14:05 <zelgomer    > assembler
2023-10-22 14:14:24 <rendar      > i see
2023-10-22 14:18:01 <rendar      > call me extremist, but i really do think if the universe we live in has been 'programmed' someway, it would be in something Forth-like :)
2023-10-22 14:26:23 <zelgomer    > i don't think it would be a discrete processor the way we imagine our processors
2023-10-22 14:34:41 <KipIngram   > I agree with that - I'm pretty much 100% convinced that no computer of the deterministic sort we build will ever house any "self-aware consciousness."
2023-10-22 14:34:50 <KipIngram   > So our universe must have something more to it than pure mechanics.
2023-10-22 14:34:56 <thrig       > Greeks thought it was a cave (they spent a lot of time in caves for some of their mystery sects)
2023-10-22 14:35:02 <KipIngram   > I can't prove that, of course.
2023-10-22 14:35:36 <KipIngram   > So I'm not on board with the idea that an AI is suddenly going to wake up and want to take us out.
2023-10-22 14:35:50 <KipIngram   > Which doesn't mean I don't think we could cause all kinds of grief for ourselves with AI.
2023-10-22 14:36:27 <KipIngram   > We'll do SOMETHING with it we shouldn't, just like we do with everything else.
2023-10-22 14:37:10 <KipIngram   > I already read a couple of years ago that some state in the US was using an AI to select criminal penalties.  I didn't like that at all.
2023-10-22 14:37:11 <thrig       > Prometheus? never...
2023-10-22 14:38:42 <KipIngram   > What was really scary about it was that when a defendant's attorney asked to inspect the AI source code, that was denied.
2023-10-22 14:39:04 <GeDaMo      > "AI is sending people to jail—and getting it wrong" https://www.technologyreview.com/2019/01/21/137783/algorithms-criminal-justice-ai/
2023-10-22 14:39:14 <thrig       > shocking
2023-10-22 14:39:24 <KipIngram   > Too easy for us to abdicate our responsibility as humans that way.
2023-10-22 14:39:49 <zelgomer    > yes, the danger with ai isn't ai itself, it's that morons who don't understand it will abuse it
2023-10-22 14:39:58 <KipIngram   > Exactly.
2023-10-22 14:40:11 <KipIngram   > We'll do plenty of amazingly good things with it too.
2023-10-22 14:40:30 <zelgomer    > looking forward to the pornos
2023-10-22 14:40:37 <KipIngram   > lmao...
2023-10-22 14:40:51 <KipIngram   > That's another thing we're guaranteed to do.
2023-10-22 14:40:53 <GeDaMo      > I think there's a subreddit for that :P
2023-10-22 14:43:30 <KipIngram   > I wouldn't object to jurors having access to AI "guidance" on setting sentences.  After all, we send random people in to be jurors.  But the final decision should be made by *people*.
2023-10-22 14:43:40 <KipIngram   > And the AI should be wide open for anyone to study.
2023-10-22 14:44:21 <KipIngram   > What will we do next?  Robot executioners?
2023-10-22 14:45:40 <KipIngram   > AI jurors?  I mean, think of the benefits - no petty human emotionalism, complete consistency, ...
2023-10-22 14:46:45 <KipIngram   > Hey - there's a science fiction story.
2023-10-22 14:47:30 <thrig       > don't need AI to have a DWB verdict
2023-10-22 14:47:30 <zelgomer    > but you're falling for the same trap now. free of human emotions and complete consistency is the polar opposite of what they're pushing out as "AI" today
2023-10-22 14:47:53 <KipIngram   > I was just adlibbing some of the arguments that might get made.
2023-10-22 14:48:01 <KipIngram   > Definitely was not advocating that.
2023-10-22 14:48:12 <zelgomer    > i see
2023-10-22 14:48:40 <KipIngram   > Actually I'm opposed to robot soldiers for the same reason - it's always possible for human armies to decide not to obey orders, if the leadership goes off the rails.
2023-10-22 14:48:46 <KipIngram   > But a robot army wouldn't, ever.
2023-10-22 14:49:05 <thrig       > unless robot army hacked
2023-10-22 14:49:10 <KipIngram   > Good point.
2023-10-22 14:49:13 <thrig       > whoops!
2023-10-22 14:50:00 <KipIngram   > And some of those Boston Dynamics videos make it seem like we are definitely moving toward the possibility of robot solidiers.  I saw one where the robot disarmed his mock attacker and then went running off down the road at like 25-30 mph.
2023-10-22 14:50:47 <KipIngram   > Probably had like a six minute battery life or somethin like that, though.
2023-10-22 14:52:26 <KipIngram   > https://www.youtube.com/watch?v=y3RIHnK0_NE
2023-10-22 14:54:22 <KipIngram   > That may not even be real, I suppose.
2023-10-22 14:54:43 <KipIngram   > Some of the commenters peg it as CGI, but if so it fooled me.
2023-10-22 14:55:40 <GeDaMo      > It's not real
2023-10-22 14:55:55 <KipIngram   > Yeah, I'm reading more closely now.
2023-10-22 14:56:16 <KipIngram   > I have seen some real Boston dynamics videos, though, and I guess that's where they're trying to go.
2023-10-22 14:56:32 <KipIngram   > I bought it initially, since i've never really paid attention to their exact progress.
2023-10-22 14:56:44 <ghoulpine   > lol "Bosstown Dynamics"
2023-10-22 14:57:03 <KipIngram   > https://bostondynamics.com/
2023-10-22 15:17:47 <KipIngram   > I don't know to what extent you guys are familiar with "Dust" short films.  There's a ton of them.  There's one I saw some years ago I'd like to watch again, but I don't know how to find it.  If anyone happens to know of an index or summary etc. of them, that would be great.
2023-10-22 15:24:58 <GeDaMo      > This is the best I can find https://letterboxd.com/hugonebula/list/dust-short-sci-fi-films/
2023-10-22 15:26:21 <GeDaMo      > https://www.themoviedb.org/video/channel/UC7sDT8jZ76VLV1u__krUutA
2023-10-22 15:39:05 <KipIngram   > Yeah, I found several such lists on letterboxd, but none of the cover images cued me in successfully.
2023-10-22 15:39:16 <KipIngram   > short of just clickin through all of them I don't see how to use that.
2023-10-22 15:39:38 <KipIngram   > What would be idea would be if there was a full list of plot summaries online somewhere, that I could do word/phrase searches on.
2023-10-22 15:40:06 <KipIngram   > One I do happen to remember the name of, though, that was also quite good and well made, was called "Plurality."
2023-10-22 15:40:25 <KipIngram   > Fun way to spend 15-20 minutes.
2023-10-22 15:50:41 <KipIngram   > I sent an email to contact@watchdust.com; maybe they have some kind of list.  You'd think they would.  Then at least I'd know I was looking at a FULL list instead of some random person't favorites.
2023-10-22 15:51:25 <KipIngram   > If anyone happens to watch "Plurality" let me know what you think.