2026-04-28 10:14:07 <veltas      > end.fs:18: Interpreting a compile-only word
2026-04-28 10:14:10 <veltas      > : TEST   >>>IF<<< ." hi" ELSE ." bye" END ;
2026-04-28 10:14:22 <veltas      > Erm... gforth you're drunk
2026-04-28 10:14:59 <veltas      > I thought compile-only words were meant to resolve the pitfalls of state-smartness
2026-04-28 10:15:32 <veltas      > This is worse
2026-04-28 10:20:43 <veltas      > This is the problem with modern Forths, they're too smart for their own good
2026-04-28 10:21:01 <veltas      > If I do something dumb I want it to crash, or silently accept it, not error-out
2026-04-28 10:21:22 <veltas      > Because maybe I want to play smart too occasionally and if it works or fails is down to me
2026-04-28 10:25:47 <tpbsd       > I agree
2026-04-28 10:26:10 <tpbsd       > crashing is a good thing (tm) in embedded Forth
2026-04-28 10:26:38 <tpbsd       > if it doesnt crash, smoke may come out of some device
2026-04-28 10:32:13 <veltas      > I'm trying to write a library to rejig a standard forth so you can use 'END' as the end of any structured syntax
2026-04-28 10:32:29 <veltas      > It should be possible in the standard, but it doesn't work in gforth 0.7.3 so it's already basically pointless
2026-04-28 10:32:53 <veltas      > And I've given up trying to 'force' gforth to do it properly for now
2026-04-28 10:34:24 <veltas      > The idea is to put the xt of the end syntax word (e.g. THEN for IF) on the stack after IF and then END just executes the xt off the stack
2026-04-28 10:35:10 <veltas      > I guess it's easier to do this for BEGIN-STRUCTURE and END-STRUCTURE as I don't think they're immediate
2026-04-28 10:35:28 <veltas      > This is the inspiration because I hate those words, it should be e.g. STRUCTURE END
2026-04-28 10:36:05 <veltas      > And END just executing the stack makes sense because it would always terminate some structured syntax, the only question is what 'action' to take at the end of the syntax
2026-04-28 11:59:37 <tpbsd       > on ##electronics a poster was having trouble with his moped battery charger using  circuit.js : https://is.gd/iu0yRh
2026-04-28 12:00:06 <tpbsd       > I've redesigned it with my system: https://raw.githubusercontent.com/techman00172/schematics/refs/heads/main/moped-batt-chgr.svg
2026-04-28 12:11:10 <veltas      > Very nice
2026-04-28 12:13:12 <tpbsd       > I've had experience with small motorbike magneto typr battery charging systems before
2026-04-28 12:14:25 <tpbsd       > qucs is shaping up quite nicely. While xschem seems far more capable, it's also a lot more complex, and I havent started watching videos on using it yet
2026-04-28 12:15:08 <tpbsd       > plux xschem uses the ancient unix widgets library
2026-04-28 12:15:32 <tpbsd       > which is ok, a blast from the past tho
2026-04-28 12:17:06 <veltas      > I can't help look at SVGs and think most SVGs generated by a program (that's not literally an SVG editor) would be smaller if they could use JS to generate the content
2026-04-28 12:17:53 <veltas      > Like the code to generate that, plus the building blocks, would surely be a lot smaller than the massive thing it generated where every single element has style info baked in
2026-04-28 12:18:12 <tpbsd       > that would mean it needs java which is not small
2026-04-28 12:18:26 <veltas      > I get why they didn't "do it properly" ... because it's hard to generate it hierarchically when your input format isn't hierarchical
2026-04-28 12:18:48 <veltas      > The whole concept of these hierarchical formats is just very flawed
2026-04-28 12:19:19 <tpbsd       > and you can see how poorly a java based sim presents with circuit.js
2026-04-28 12:19:32 <veltas      > That's apples and oranges
2026-04-28 12:20:43 <tpbsd       > I see people everyday using circuit.js and it's so bad that they are wasting days with their design because of it, so it's more like sulphuric acid vs lemonade to me
2026-04-28 12:21:45 <veltas      > That's got nothing to do with what I was talking about though, I am saying it would be more compact to generate the SVG elements with JS than to just list them all fully without any hierarchy, as your app has done
2026-04-28 12:21:51 <tpbsd       > as a electronics guy, I'm only interested in the end result from a pov of how it helps the electronics guy, not how efficient the programming is
2026-04-28 12:22:00 <veltas      > This isn't a major problem but it just is a bit sad because vector formats should be small
2026-04-28 12:22:07 <veltas      > But in practice SVGs are often very large
2026-04-28 12:22:22 <tpbsd       > youre coming from a programmer perspective tho, I'm from 180 degrees away
2026-04-28 12:22:37 <veltas      > I'm not saying it's a problem, just saying it's sad
2026-04-28 12:22:39 <tpbsd       > I agree, they should be small
2026-04-28 12:22:49 <veltas      > That the engineering has let the user down
2026-04-28 12:23:18 <tpbsd       > some 733t programmer should fix that for qucs :-)
2026-04-28 12:23:43 <tpbsd       > maybe someone will one day
2026-04-28 12:23:59 <tpbsd       > but with the rise of AI, I'm not holding my breath
2026-04-28 12:24:45 <tpbsd       > after all of us baby boomers are dead, and your generation also, this world will forget how to program
2026-04-28 12:25:23 <tpbsd       > the skill will be gone, just like basket weaving
2026-04-28 12:25:35 <veltas      > I'm still not convinced AI will remove programmers
2026-04-28 12:25:55 <tpbsd       > Ai wont, managers and accountants will
2026-04-28 12:25:58 <veltas      > Yeah
2026-04-28 12:26:05 <tpbsd       > anything to lower costs
2026-04-28 12:26:05 <veltas      > It can probably speed up the decline of a lot of software
2026-04-28 12:26:09 <tpbsd       > yeah
2026-04-28 12:26:35 <tpbsd       > I hear theyre already using AI to offer programmer hires less $$
2026-04-28 12:26:42 <veltas      > I'm very impressed by AI and think it's huge, but I really don't think 'vibecoding' etc is a good idea for any serious kind of programming product
2026-04-28 12:27:08 <tpbsd       > as in. we can just use a trainee and AI, why do we need you ?
2026-04-28 12:27:28 <tpbsd       > I share your opinion
2026-04-28 12:27:48 <veltas      > The principles haven't changed in programming at all, you still end up maintaining a program you'd rather was closer to a well documented and precise spec than just spaghetti code
2026-04-28 12:27:54 <veltas      > AI is just the new spaghetti
2026-04-28 12:28:18 <veltas      > It looks better than the old, but experience will shine a light on it
2026-04-28 12:28:53 <tpbsd       > ok, I vibe coded 2 apps, maybe 6 months ago, but I didnt enjoy the journey as I felt blindfolded, and I dont know anything about those apps (python and lua)
2026-04-28 12:29:11 <veltas      > I think people firmly rejecting AI like a luddite will do better than people firmly embracing it and vibing everything... unless your goal is to get venture capital funding and not actually to create something
2026-04-28 12:29:44 <tpbsd       > I love AI as a superior Google replacement, but I'll never use it to code with again
2026-04-28 12:29:46 <veltas      > Then you can maybe make a juicy salary before it all falls apart and buy a big mansion or whatever, but that stuff doesn't interest me because I'm an engineer
2026-04-28 12:30:28 <tpbsd       > and I love Forth coding and design, I dont eant a AI to do that
2026-04-28 12:30:34 <tpbsd       > want
2026-04-28 12:30:58 <tpbsd       > that would be like buying a icecream and getting AI to eat it
2026-04-28 12:31:44 <tpbsd       > plus,in using my AI as a superior google, I find it makes too many mistakes to ever rely on it for design
2026-04-28 12:32:09 <veltas      > Just saw someone complain that the book I'm reading, published in 1975, is called "The Mythical Man-Month", i.e. it has 'Man' rather than 'person' in the title
2026-04-28 12:32:16 <veltas      > What a highly ignorant thing to complain about
2026-04-28 12:32:36 <veltas      > "Man" isn't gendered there, it could mean a man or a woman
2026-04-28 12:32:50 <veltas      > It might not be typical today to write that, but this was written in 1975
2026-04-28 12:35:03 <veltas      > I'm glad I don't think that way and can actually just read and enjoy this brilliant work, rather than worrying about such things
2026-04-28 12:35:09 <tpbsd       > it's a mad world, they should be reading the book, not losing their mind over the title
2026-04-28 12:35:56 <tpbsd       > changing that title, is a sneaky way to remove our past and our traditions
2026-04-28 12:36:20 <veltas      > Well also it would ruin the alliteration
2026-04-28 12:36:33 <tpbsd       > people like that may as well be ISIS destroying 5000 year old religous sites
2026-04-28 12:37:08 <veltas      > You'd have to call it the "Phony People-Period" or something
2026-04-28 12:37:11 <tpbsd       > or blowing up awesome stone statues made 6000 years ago
2026-04-28 12:37:34 <veltas      > Yeah that really tells the whole world in actions rather than words exactly what kind of people they are
2026-04-28 12:37:42 <tpbsd       > yeah
2026-04-28 12:37:59 <tpbsd       > they want a new world, and we arent in it
2026-04-28 12:38:26 <veltas      > Rather I would say they're ignorant and destructive
2026-04-28 12:38:57 <tpbsd       > these are the same mental midgits that passed a law making the value of PI = 3.2 because the real value is too unwieldy
2026-04-28 12:39:50 <tpbsd       > and to imagine that I actually believed, up until recently, that Nero actually owned and played a fiddle
2026-04-28 12:57:41 <veltas      > My i9 desktop at work feels more sluggish than my Core 2 Duo thinkpad at home
2026-04-28 12:57:46 <veltas      > That's the power of Windows 11
2026-04-28 13:00:35 <tpbsd       > feel the POWARGH!
2026-04-28 13:13:52 <vms14       > tpbsd: about ai, i also wonder about people vibecoding
2026-04-28 13:14:02 <vms14       > but i use free models so they are limited
2026-04-28 13:14:10 <vms14       > seems if you pay they get better
2026-04-28 13:14:30 <vms14       > i only use it as a google search, but also to write emacs conf and help with css
2026-04-28 13:14:38 <vms14       > and setup stuff like git xd
2026-04-28 13:15:06 <vms14       > if i for example give the interpreter code it will not realize a lot of stuff
2026-04-28 13:15:17 <KipIngram   > I'm just fundamentally opposed to the whole idea of vibe coding.
2026-04-28 13:15:41 <vms14       > well i would not accept code from random people unless it's a library i have to use
2026-04-28 13:15:42 <tpbsd       > KipIngram, so am I
2026-04-28 13:16:00 <vms14       > even less from a bot that does not even understand my code
2026-04-28 13:16:02 <tpbsd       > I tried it, made stuff, but didnt like it at all
2026-04-28 13:16:23 <KipIngram   > I think it's extremely dangerous to just trust the output of an AI.  My feeling is that it's quite fine to use it as a sounding board and that kind of thing, but once you're using it to produce stuff that you don't yourself understand you're in dangerous territory.
2026-04-28 13:16:31 <vms14       > it helps me with css though, i come crying to chatgpt when my divs won't center xd
2026-04-28 13:16:47 <KipIngram   > It's quite clear to me, though, that lots and lots of people will try to use it to "look better than they really are."
2026-04-28 13:17:01 <vms14       > i think they are smoke and we project the rest
2026-04-28 13:17:14 <vms14       > but it seems that latest models can do impressive stuff
2026-04-28 13:17:25 <KipIngram   > It's just human nature to try to accomplish as much as you can in spite of your limitations.
2026-04-28 13:18:37 <KipIngram   > vms14: It's harmless enough when  you're just plunking around on some hobby thing - my worry is that people doing serious work that affect safety and health will start going down that road.
2026-04-28 13:18:49 <KipIngram   > affects
2026-04-28 13:19:24 <tpbsd       > agreed
2026-04-28 13:19:30 <KipIngram   > Whether your div is centered or not doesn't really fall into that category.
2026-04-28 13:19:37 <vms14       > yeah we will add more and more responsibility on ai
2026-04-28 13:19:42 <tpbsd       > but vibe coding will kill innovation imho
2026-04-28 13:19:59 <vms14       > i love programming, so if you write the code for me what i have left
2026-04-28 13:20:04 <tpbsd       > and programmers become clerks with an AI
2026-04-28 13:20:08 <KipIngram   > Right - I agree.  The "sounding board" use doesn't, though - it might even enhance it.
2026-04-28 13:20:09 <vms14       > i don't care about the app, but the process of making it
2026-04-28 13:20:19 <tpbsd       > vms14, +1
2026-04-28 13:20:31 <vms14       > i would love if i could use ai to help me reason about stuff, but it's quite limited there
2026-04-28 13:20:52 <vms14       > i feel sad because it has so much information but does not apply it properly
2026-04-28 13:21:07 <vms14       > with some specific instructions i might get somewhere but meh
2026-04-28 13:21:42 <KipIngram   > vms14: My impression is that it's very prone to using deprecated information about things like software tools.
2026-04-28 13:24:28 <KipIngram   > css suffers the same issue that LaTeX does.  It's not a "smoothly architected tool" that runs on "principles."  It's just a hodge podge of little bits all thrown together in a blender.
2026-04-28 13:25:06 <vms14       > well my problem is underestimating css and not willing to spend time learning it properly
2026-04-28 13:25:09 <tpbsd       > I have a updated https://raw.githubusercontent.com/techman00172/schematics/refs/heads/main/moped-batt-chgr.svg
2026-04-28 13:25:49 <tpbsd       > SPICE seems to think that 'ordinary' diodes conduct both ways in some cases
2026-04-28 13:26:10 <tpbsd       > so it's not only AI that screws up
2026-04-28 13:27:08 <KipIngram   > My PERSONAL problem with css is also the same as with LaTeX - I don't need to use it often enough to retain it well.
2026-04-28 13:27:33 <KipIngram   > When I need one of them, I need it - but then it might be a year or two before I need it again.
2026-04-28 13:27:47 <KipIngram   > And I'm basically starting over.
2026-04-28 13:28:21 <tpbsd       > use it or lose it ?
2026-04-28 13:28:29 <KipIngram   > Yeah.
2026-04-28 13:28:47 <KipIngram   > "Occasional use" + "no organizing principles" = "hard to learn."
2026-04-28 13:29:16 <veltas      > vms14: I'm hearing from people that the models you pay for aren't that much better, they all have similar hangups
2026-04-28 13:29:22 <tpbsd       > Im the same with programming, I have to do it daily or I forget the important stuff
2026-04-28 13:29:51 <vms14       > veltas: i see people using claude and codex
2026-04-28 13:30:55 <tpbsd       > we are reliving history! Nero is playing his fiddle while Rome burns, and the king has amazing new clothes !
2026-04-28 13:33:32 <KipIngram   > It's a weird time.  The US is having one of it's "every 2-3 generations" political resets.  Hard to know what it will look like when the dust settles and the new lines solidify for a while.
2026-04-28 13:33:55 <KipIngram   > I think Trump is more of a symptom than a cause.
2026-04-28 13:34:31 <tpbsd       > absolutely
2026-04-28 13:37:10 <veltas      > So tpbsd I just received a cheap chinese radio as a present, it comes with a UK power adaptor which seems to just be the plug (a small box) and claims it's 5V DC output (with a USB-C plug at other end which it warns you not to plug into anything else)
2026-04-28 13:37:37 <veltas      > So what sort of circuit would you expect in the mains plug?
2026-04-28 13:37:46 <tpbsd       > I have a Chinese radio and it's awesome!
2026-04-28 13:38:04 <KipIngram   > If it's really 5V DC output, then you should be able to plug it into anything.
2026-04-28 13:38:11 <veltas      > Mine's alright so far, it's got FM/DAB/DAB+.  The controls were a bit weird though.
2026-04-28 13:38:20 <tpbsd       > veltas, probably a mains switching supply, theyre tiny now
2026-04-28 13:38:38 <tpbsd       > veltas, what is the model ?
2026-04-28 13:38:44 <tpbsd       > Tecsun ?
2026-04-28 13:39:07 <tpbsd       > mine has that plus short and long wave
2026-04-28 13:39:10 <KipIngram   > veltas: I've got a little USB dongle gadget that's a voltage/current meter.  Tells you exactly what's going on.
2026-04-28 13:39:28 <veltas      > tpbsd: I would like short + long wave but didn't see any cheap options that supported that
2026-04-28 13:39:36 <veltas      > Which is bizarre
2026-04-28 13:39:43 <veltas      > Should be easiest to decode
2026-04-28 13:39:54 <veltas      > But I guess it's more confusion in the menu
2026-04-28 13:40:13 <KipIngram   > tpbsd: Do you actually hear much on shortwave?  When I was a kid I was pretty into SWL - I found the stuff out there fascinating.  Then I got busy with college and so on and didn't listen for a long time.  When I went back the whole thing seemed dead by comparison to the old days.
2026-04-28 13:40:29 <KipIngram   > Most countries have shut down their shortwave broadcasting - the internet sort of killed it.
2026-04-28 13:40:44 <KipIngram   > Mostly what I hear out there these days is fringe religious broadcasters.
2026-04-28 13:40:47 <veltas      > tpbsd: "AZATOM Desire X2"
2026-04-28 13:41:25 <KipIngram   > I think closing down, or greatly reducing, things like Voice of America, the BBC, and so on was a mistake.
2026-04-28 13:41:39 <KipIngram   > Enemy nations can control their internet, but you can't stop shortwave signals.
2026-04-28 13:42:18 <KipIngram   > Surely those things weren't THAT expensive compared to the whole national budget.
2026-04-28 13:42:24 <tpbsd       > KipIngram, yeah with a long antenna
2026-04-28 13:42:54 <KipIngram   > I didn't really mean "can you get signals" - I was more asking if you still find worthwhile content.
2026-04-28 13:43:01 <tpbsd       > KipIngram, there are TONS of shortwave stations world wide still
2026-04-28 13:43:32 <veltas      > Long wave is the only radio station that actually has decent signal the whole drive home
2026-04-28 13:43:40 <veltas      > BBC Radio 4 which is good sometimes
2026-04-28 13:43:40 <KipIngram   > Hmmm.  Maybe I'll try again.  I have to admit my later antenna efforts were pretty half-assed.  Hard to have a good antenna when you live in an apartment.
2026-04-28 13:43:44 <tpbsd       > well there is the language barrier as I only speak english, and a lot are not english
2026-04-28 13:44:07 <KipIngram   > I'm English only as well, which I feel a little bad about.
2026-04-28 13:44:23 <tpbsd       > but yeah, it's always enlightening to get a non mainstream media perspective
2026-04-28 13:44:27 <KipIngram   > I took French in high school, but it wasn't very demanding and even then the best I could do was read a little of it.
2026-04-28 13:44:47 <veltas      > I thought French in the UK was redundant but in the USA?
2026-04-28 13:45:15 <veltas      > I guess in Canada they speak French
2026-04-28 13:45:21 <KipIngram   > I think my little school just wanted to offer a language and only had someone capable of teaching French.
2026-04-28 13:45:32 <KipIngram   > I live in Texas now, so Spanish would really be the right language to learn.
2026-04-28 13:45:39 <veltas      > If I was american I'd probably want to learn Spanish
2026-04-28 13:45:52 <KipIngram   > The one that might actually help me out in the world.
2026-04-28 13:45:53 <veltas      > Given so many people speak spanish as primary or second lang in US
2026-04-28 13:45:55 <veltas      > Seems practical
2026-04-28 13:46:06 <KipIngram   > Yes - it would be the most useful one.
2026-04-28 13:46:22 <KipIngram   > I wish I could read German - there are a lot of classic science papers that were done in German.
2026-04-28 13:46:33 <KipIngram   > In the run-up to World War I Germany really dominated science.
2026-04-28 13:47:20 <KipIngram   > My dad taught himself German in graduate school for that reason - he was an organic chemist and needed to read some of those papers.
2026-04-28 13:47:30 <KipIngram   > He definitely didn't "speak it," though.
2026-04-28 13:47:53 <KipIngram   > Beyond a handful of words he remembered.
2026-04-28 13:48:16 <KipIngram   > I'd also like to be able to read Latin and ancient Greek, for the history value.
2026-04-28 13:48:46 <tpbsd       > all good ideas
2026-04-28 13:49:04 <KipIngram   > We used to teach people Latin, before the education system crumbled away.
2026-04-28 13:50:54 <KipIngram   > I did make a run at Latin a while back, and may again - there's a set of books out there designed for you just to "learn it by reading it."  I've got one of them on my coffee table now - it's called "Familia Romana," and the idea is that you just start on page one and start reading.  It's very simple, and illustrated, so you can pretty much get from context what it's saying.
2026-04-28 13:51:01 <KipIngram   > Then as you work through it it gets slowly more advanced.
2026-04-28 13:51:30 <KipIngram   > Starts off with geography, family relationships, and so on.
2026-04-28 13:52:03 <KipIngram   > I also feel like learning Latin first would make learning the later Romance languages easier too.
2026-04-28 13:52:48 <KipIngram   > The whole business of linguistics is pretty fascinating to me too - apparently languages evolve in standard, sometimes predictable ways.
2026-04-28 13:53:12 <KipIngram   > They can map those changes and use them similar to the way that use DNA scans to map migration around the world over history.
2026-04-28 13:53:51 <veltas      > tpbsd: https://www.azatom.com/power-adapter-for-desire-dab-radio-speaker-desire-x
2026-04-28 13:54:02 <KipIngram   > Apparently most of the languages we speak today evolved from a common ancestor they call Proto Indo European, which developed around the Ukraine / Russian steppes region.
2026-04-28 13:54:52 <KipIngram   > Even Sanskrit ties back to PIE.
2026-04-28 13:55:52 <veltas      > When I first met my Indian friend's wife, I looked her name up "Karishma" because I like etymology.  It really derives from the root of the word "Charisma".  I didn't expect it to be so close to an English word
2026-04-28 13:56:05 <veltas      > Although the meaning and history is not what we think of re "Charisma" nowadays
2026-04-28 13:57:13 <KipIngram   > Yeah.  Also "Kaiser" and "Czar" both derive from Caesar.
2026-04-28 13:57:51 <veltas      > I think we mention that once a week in here
2026-04-28 13:58:06 <KipIngram   > :-)  Really?  That's one I've missed somehow.
2026-04-28 13:58:21 <veltas      > It's actually quite mad just how much we've mentioned the Roman Empire today, and yet quite normal
2026-04-28 13:58:43 <KipIngram   > Some Portuguese guy was the first to notice similarities with the languages down in India with the western European languages.
2026-04-28 13:59:06 <veltas      > I doubt he was the first
2026-04-28 13:59:13 <tpbsd       > veltas, any USB type C charger would work I think
2026-04-28 13:59:16 <KipIngram   > Okay, he popularized it.
2026-04-28 13:59:30 <KipIngram   > In intellectual circles.
2026-04-28 14:01:16 <KipIngram   > veltas: Apparently there's a pop folklore tidbit that claims a lot of men particularly think about the Roman empire several times a day.  In the West, at least.
2026-04-28 14:01:43 <veltas      > I would say if the middle east was more stable we would be citing some Iraqi guy or similar, alas we are not
2026-04-28 14:01:44 <KipIngram   > I don't really think I do unless I happen to be looking at that part of history, but who knows?
2026-04-28 14:02:09 <veltas      > That's more of a meme but it's a funny and wholesome meme
2026-04-28 14:02:14 <KipIngram   > Oh yeah - Iran and Iraq were really at the very beginning.
2026-04-28 14:02:30 <KipIngram   > Yes - meme is the right word.
2026-04-28 14:03:41 <veltas      > The meme's more along the lines of women asking men "do you really think about the Roman empire every day" and all men replying "yes"
2026-04-28 14:03:58 <veltas      > Or thinking about it a bit and saying "I would have to say yes"
2026-04-28 14:04:14 <KipIngram   > A day or so ago I found myself thinking that a good "alternate history novel" might be built around the idea that Carthage managed to not get obliterated by the Romans and developed into a successful rival power going into more modern times.
2026-04-28 14:04:23 <veltas      > There is definitely some gender humour in there somewhere
2026-04-28 14:04:47 <KipIngram   > Yeah, that's the context I've heard the meme in - women talking about men.
2026-04-28 14:05:41 <KipIngram   > There was some pretty funny Rome humor in The Sopranos, too - Tony definitely fancied himself a modern Roman centurion.
2026-04-28 14:05:58 <KipIngram   > And that show had elements of his "psychological activity" woven into it.
2026-04-28 14:06:51 <veltas      > The Sopranos is my favourite TV show
2026-04-28 14:07:23 <tpbsd       > I'ne never seen it but read rave reviews
2026-04-28 14:07:30 <veltas      > Ugh need to get a DVD player so I can watch my Sopranos boxsets again
2026-04-28 14:07:45 <KipIngram   > It's pretty good.  Like most good shows, it's got ups and downs.
2026-04-28 14:08:01 <veltas      > I think it gets worse towards the end, but it's so good that I forgive it
2026-04-28 14:08:28 <KipIngram   > Right.  That's a bit how I feel about Mad Men too.  It definitely rolled off in later seasons, but seasons 1 and 2 are really pretty fantastic.
2026-04-28 14:08:30 <veltas      > TV usually gets worse before it ends, I can't think of a lot of counter-examples
2026-04-28 14:08:57 <KipIngram   > Well, a) they run out of ideas, and b) they just want to keep making money beyond the stretch of the story.
2026-04-28 14:09:32 <veltas      > You need to finish Mad Men
2026-04-28 14:09:58 <KipIngram   > I have seen almost all of it.  I just enjoy the first couple of seasons the most.
2026-04-28 14:10:04 <veltas      > But like whisky Mad Men needs to be sipped, not gulped
2026-04-28 14:10:12 <KipIngram   > Loved the way it ended.
2026-04-28 14:10:34 <KipIngram   > Such a nice tie-back to the "It's TOASTED" moment in the very first episode.
2026-04-28 14:10:59 <KipIngram   > I was just FASCINATED with the Don Draper character.
2026-04-28 14:11:20 <veltas      > That show got me into whisky and I don't regret it
2026-04-28 14:11:52 <veltas      > Although I'm not really 'drinking' but I can drink whisky and enjoy it if I ever needed to
2026-04-28 14:12:09 <KipIngram   > The small company environment of Sterling Cooper appealed to me wildly - their business was in no way "my field," but atmosphere-wise I'd have thrived in an environment like that.
2026-04-28 14:12:16 <veltas      > I also tried taking up smoking but that's a little too expensive for me, and my wife won't let me
2026-04-28 14:12:22 <KipIngram   > LOVED Burt Cooper too.
2026-04-28 14:12:29 <veltas      > See I need to be married because my wife literally keeps me out of trouble
2026-04-28 14:12:39 <KipIngram   > I think that's true for most men.
2026-04-28 14:13:08 <veltas      > Burt Cooper is meant to be a Randian villain but you don't see a lot of the downsides of that until the last series
2026-04-28 14:13:51 <KipIngram   > He won my heart with "Who cares?"
2026-04-28 14:13:59 <KipIngram   > THAT was a great scene.
2026-04-28 14:14:06 <veltas      > I think Ayn Rand is cool, let the haters hate.  I mean I am not Randian but she knew that communism sucked and helped a lot of people get over that
2026-04-28 14:14:16 <veltas      > Ayn Rand was like communism shock-therapy
2026-04-28 14:14:19 <KipIngram   > I feel the same way.
2026-04-28 14:14:21 <veltas      > I think it had its place
2026-04-28 14:14:36 <KipIngram   > Was she entirely right?  No, but at the same time she saw something important.
2026-04-28 14:14:47 <veltas      > And it was based on her lived experiences too
2026-04-28 14:14:59 <KipIngram   > She clearly recognized the "anti-individual" nature of communism.
2026-04-28 14:15:11 <veltas      > She knew communism was evil and had a pretty interesting albeit spicy way to explain it
2026-04-28 14:15:19 <KipIngram   > I have no interest in living in an ant colony.
2026-04-28 14:15:47 <veltas      > Or next to an ant colony, based on what happened to the ant colony in mad men
2026-04-28 14:16:09 <KipIngram   > Uh oh - that one has slipped my mind.
2026-04-28 14:16:40 <KipIngram   > My whole world view, if it's based on anything, is based on the "primacy of the individual."
2026-04-28 14:17:43 <KipIngram   > But the closing scene of the Mad Men pilot episode nearly knocked me out of my chair.  What a twist.
2026-04-28 14:18:00 <veltas      > My favourite bit was "it's toasted"
2026-04-28 14:18:17 <KipIngram   > Oh yeah - that's where you first saw Don's "peculiar genius."
2026-04-28 14:18:20 <veltas      > But yeah they ended strong
2026-04-28 14:18:39 <KipIngram   > He was just plugged into "everyman" in a way that the guys around him just couldn't compete with.
2026-04-28 14:19:18 <KipIngram   > He just seemed to "get" what was going to resonate emotionally with people.
2026-04-28 14:20:36 <KipIngram   > Meanwhile he had some sort of "empty spot" inside himself that he was desperately trying to fill and couldn't.  I think the opening intro sequence, with the guy falling, captures that really well.
2026-04-28 14:32:10 <veltas      > Probably the best bit is when they split from Sterling Cooper
2026-04-28 14:32:36 <veltas      > I feel like it would have been interesting to see more of the operating-out-of-a-hotel stage but maybe I'm wrong
2026-04-28 14:33:38 <KipIngram   > That bit was really fun.  And the way they accomplished it contractually was pretty awesome too.
2026-04-28 14:34:07 <KipIngram   > Oh, and I can't let myself depart from a Mad Men conversation without noting that I LOVED - absolutely loved - Joan.  So sexy.
2026-04-28 14:50:45 <veltas      > tpbsd: Is there a safe way to test whether it's just standard USB-C power?
2026-04-28 14:53:31 <veltas      > KipIngram: What do you reckon the 'simplest' model for Forth is on AMD64, ITC, DTC, tokens, or STC?
2026-04-28 14:53:54 <veltas      > Assuming also that machine code is kept separate to the main dictionary
2026-04-28 14:54:57 <veltas      > I think ITC might fit the bill, or DTC with reverse references to dictionary
2026-04-28 14:55:37 <veltas      > STC has that advantage that a lot of stuff is just automatic if you use the native execution model, but many disadvantages like can't use push/pop
2026-04-28 15:02:43 <KipIngram   > Hmmm.  I'm not sure - I've always been a big fan of ITC.  I think particularly it makes it easier to implement CREATE/DOES>.
2026-04-28 15:02:56 <KipIngram   > I definitely think it has the most "grace."
2026-04-28 15:04:44 <veltas      > Yeah ITC is smaller than DTC and simpler for CREATE/DOES>
2026-04-28 15:06:54 <KipIngram   > I've only written indirect and direct threaded systems - haven't ever done a code threaded one yet, though I regard the one I'm about to do as code-threaded.  VM code, but "the Forth" will still be code threaded.
2026-04-28 15:07:50 <KipIngram   > You more or less CAN'T completely segregate your code in a direct threaded system.
2026-04-28 15:08:01 <veltas      > When I hear "VM" I think "tokenised but not in assembly"
2026-04-28 15:08:05 <KipIngram   > Since every word has to have a code snip at its CFA.
2026-04-28 15:08:31 <KipIngram   > Pardon my nomenclature - I'm planning to implement a thin VM layer and then will use that layer to write a code-threaded Forth.
2026-04-28 15:08:40 <KipIngram   > The VM layer will be byte-coded.
2026-04-28 15:09:04 <veltas      > No your nomenclature is fine
2026-04-28 15:09:09 <KipIngram   > The words you'd normally expect to be "primitives" will largely be vm instructions instead.\
2026-04-28 15:09:43 <veltas      > I'm just saying "VM" is a way people beautify writing the kernel for a tokenised Forth in C
2026-04-28 15:10:13 <veltas      > The main difference between tokenised Forth and Vm Forth is I don't expect to be able to add CODE words (i.e. native non-VM words) in a VM forth
2026-04-28 15:11:06 <KipIngram   > I wrote one system in C, but didn't do any tokenizing - instead I just used C to create a fairly standard ITC system.  The primitives were C code snips, but they used gcc's pointer to label extension to do a "proper" machine code threading operation.
2026-04-28 15:11:21 <KipIngram   > The rest of the system, aside from those primitive snips, lived in a an array.
2026-04-28 15:12:16 <veltas      > Yeah in most high level languages you can make ITC or tokenised/VM systems easily
2026-04-28 15:12:27 <veltas      > vms14's javascript system I consider ITC
2026-04-28 15:13:04 <KipIngram   > The source code for that wound up terribly ugly - I tried to use macros to make it "cleaner," and that worked for a while, but by the time it was done it was fairly hideous.
2026-04-28 15:13:10 <veltas      > Not even that much of a stretch, although I think the actual way that the word is traversed is more of a lisp-style list than an array, so that's one difference
2026-04-28 15:13:24 <KipIngram   > It worked well, but I eventually abandoned it because the C was too hard to maintain.
2026-04-28 15:13:33 <KipIngram   > That was when I switched to nasm.
2026-04-28 15:18:35 <veltas      > Are you still using nasm?
2026-04-28 15:26:56 <KipIngram   > I'm trying to do without it this time - I'm interested in building this system from a "meta source" file that I plan to process with a Python program that uses a machine code generator library.
2026-04-28 15:27:18 <KipIngram   > The idea is for that same file to later be consumable by the Forth itself for self-recompiliation.
2026-04-28 15:27:38 <KipIngram   > So I'm designing the meta-source to be "Forth friendly," and the Python will just have to cope with that.
2026-04-28 15:28:00 <KipIngram   > Once I get it bootstrapped I won't need the Python anymore -  it'll be self-standing.
2026-04-28 15:28:30 <veltas      > I forgot if there was a reason you're not cross-compiling from an existing forth?
2026-04-28 15:29:05 <veltas      > Like you could probably jerry-rig gforth to be quite similar to how you code and get it to build your forth... although after my experience today with gforth I don't know
2026-04-28 15:29:10 <KipIngram   > This ESP32-C6 is RISCV, so I'm planning to use the riscvlib package.
2026-04-28 15:29:47 <KipIngram   > No particularly profound reason - just not how I chose to pursue it.
2026-04-28 15:31:05 <KipIngram   > It's really time to start that - I have this thing to the point where I can run arbitrary machine code that I've poked into ESP32 RAM, so now insted of doing that by hand I need to "make it from meta-source."
2026-04-28 15:32:28 <veltas      > Sounds like it would be easier to write a traditional forth, more steps to get to a VM
2026-04-28 15:32:30 <KipIngram   > Honestly I've just never really had much interest in "other people's Forths."
2026-04-28 15:32:50 <veltas      > Traditional i.e. written in an assembler
2026-04-28 15:33:44 <KipIngram   > Being able to re-compile it self-standing later is important to me, so I figure I may as well go ahead and get the source for that nailed down right from the start.
2026-04-28 15:34:02 <KipIngram   > If I write it in assembly I'll almost certainly cut corners on that somewhere.
2026-04-28 15:34:51 <KipIngram   > I think most Forths that can re-compile themselves do so via a fair bit of "hoop jumping" and manual keeping up of things - I'm trying to make it much more seamless in this design.  More "natural."
2026-04-28 15:48:30 <KipIngram   > One plan there is to introduce the notion of an "image" as a first class construct.  Then you'll be able to say "<parameter list> image <image name>", and then "<image name> target" at any time to make that image the recipient of any new compilation.  It would even be possible to have several images around at once and switch among them, but I doubt that's useful.  The parameter list will define
2026-04-28 15:48:33 <KipIngram   > things like processor architecture, cell size (and just plain old size) etc.  The "running system" will occuply image "live," so when you're done "live target" will put you back into normal operating mode.
2026-04-28 15:58:43 <veltas      > I prefer ugly hacks in the cross-compile, rather than adding more primitives to base system, personally
2026-04-28 16:23:03 <KipIngram   > Fair enough.  I like cleanliness.
2026-04-28 16:29:41 <KipIngram   > I don't think there's a real difference between the two - I'm just preferring to abstract/encapsulate some of the necessary bits.
2026-04-28 16:46:31 <KipIngram   > And in any case the actual code to manage it all will only get loaded when I want to recompile - the only "always there" cost will be for the live image to have a small bit of data stored in whatever way I decide is appropriate.  Probably just a few dozen bytes at most.
2026-04-28 16:46:54 <KipIngram   > I don't plan for all the meta-compile code to be live on system start-up.
2026-04-28 16:47:43 <KipIngram   > And the way I plan to do this I'll be able to load a vocabulary, use it, and then unload / deallocate it - it'll compile into its own RAM region that I can discard when I don't need it any more.
2026-04-28 16:48:22 <KipIngram   > The whole  idea there is to be able to "run applications" the way I can in Linux - they'll just load from source instead of from some binary pre-compiled format.
2026-04-28 16:58:22 <veltas      > Yeah agreed
2026-04-28 16:59:10 <veltas      > That's why I would do hacks on cross-compile, because I want all that stuff only built on demand, it's not something that should be present on the base system
2026-04-28 16:59:21 <veltas      > But leaving some hooks etc makes sense
2026-04-28 17:01:06 <KipIngram   > I totally agree with you that it should be "transient" functionality.  A lot of stuff I want falls into that category - for scientific computing I want to be able to have geometric algebra entities in a first class way, but that's clearly not something you'd find on start-up.
2026-04-28 17:06:33 <veltas      > Yeah I've often thought about how to make a forth system flexible with any order of starting 'apps' and unloading them
2026-04-28 17:07:08 <veltas      > But I think really Forth is never meant to be *that* convenient to the end user, at sacrifice of complexity.  The workaround is just forget it all and reload apps if they are nuked in the middle
2026-04-28 17:08:02 <veltas      > I guess you could have different areas you allocate to different 'users' (which could be one user but different apps they know they want to not impact each other)
2026-04-28 17:08:44 <veltas      > The more important challenge is just getting an "OS" with nice apps 'working' in the first place!
2026-04-28 17:11:16 <KipIngram   > No, it really wasn't intended to do that initially.
2026-04-28 17:11:36 <KipIngram   > I think Chuck clearly had in mind a system (computer and all) that had one job that it just did continuously.
2026-04-28 17:12:03 <KipIngram   > And for that, especially given that it often involved controlling equipment and what-not, Forth is virtually perfect.
2026-04-28 17:12:31 <veltas      > Honestly rebuilding is pretty fast, even with linked list on Z80
2026-04-28 17:12:55 <veltas      > So there we have a feature we have already for free, we should use that
2026-04-28 17:13:06 <veltas      > Long-term data needs to be in blocks or files already
2026-04-28 17:13:42 <veltas      > If re-building something takes a lot of input then ... put that in a block/file and run that instead
2026-04-28 17:14:02 <veltas      > Need to encourage forth users to make use of the tools they have
2026-04-28 19:23:51 <tpbsd       > veltas, if it has a USB-C connector, I think it will be USB 5V
2026-04-28 21:17:24 <veltas      > Yeah I'm just wondering about their warning in the manual "Caution: Only use the supplied power adapter for this device. Do not
2026-04-28 21:17:27 <veltas      > use for other devices"
2026-04-28 21:18:29 <tpbsd       > ahh
2026-04-28 21:18:55 <tpbsd       > maybe it isnt USB standard ?
2026-04-28 21:19:48 <tpbsd       > my radio is USB-C and any USB adaptor works fine
2026-04-28 21:20:25 <tpbsd       > a manufacturer would have to be insane to not use the USB standard ?
2026-04-28 21:20:40 <tpbsd       > besides, all the chipsets etc are made for that standard
2026-04-28 21:21:04 <veltas      > Yeah might not have a USB-C controller maybe?
2026-04-28 21:21:32 <veltas      > Needs a special controller to handle the different orientations or is it symmetrical for power?
2026-04-28 21:22:18 <tpbsd       > I dont know, I havent looked into it
2026-04-28 21:22:29 <tpbsd       > but I think their warning is bs
2026-04-28 21:22:42 <veltas      > Sorry tpbsd I just assumed you know everything about electronics lol
2026-04-28 21:23:01 <veltas      > It could be, that's why I'm wondering how to test, I guess probe the pins and see what's what
2026-04-28 21:23:06 <tpbsd       > no, only AI and my sister know everything
2026-04-28 21:23:11 <veltas      > :P
2026-04-28 21:23:42 <tpbsd       > just plug the power unit into some usb-c device and see if it works >
2026-04-28 21:23:44 <tpbsd       > ?
2026-04-28 21:24:28 <veltas      > Yeah power's symmetric
2026-04-28 21:24:51 <veltas      > VBUS and GND are rotationally symmetrical
2026-04-28 21:26:16 <veltas      > All I've got to test is my phone, so I'd rather not
2026-04-28 21:26:33 <veltas      > I'd be more inclined to plug my phone charger into the radio
2026-04-28 21:27:34 <tpbsd       > is that all ?, no usb torch ?
2026-04-28 21:27:41 <veltas      > Nope
2026-04-28 21:27:47 <tpbsd       > no usb st-link ?
2026-04-28 21:28:04 <tpbsd       > no stm32 discovery board ?
2026-04-28 21:28:26 <veltas      > I have a STM32 discovery board
2026-04-28 21:28:40 <veltas      > Can't remember if it's USB-C but I don't want to brick it
2026-04-28 21:30:32 <tpbsd       > it wont be type C anyway, and you cant brick it. The power just goes into a three term reg
2026-04-28 21:34:26 <tpbsd       > veltas, are you able to read https://raw.githubusercontent.com/techman00172/schematics/refs/heads/main/moped-batt-chgr.svg ok ?
2026-04-28 21:34:57 <tpbsd       > one user has reported that the text in the boxes overflows the box, do you see that ?
2026-04-28 21:43:19 <veltas      > Are they on a Mac?
2026-04-28 21:43:31 <veltas      > Oh actually it overflows on my laptop
2026-04-28 21:43:35 <veltas      > Looked fine on Windows in the office
2026-04-28 21:44:59 <veltas      > The font in SVG just says "Sans Serif" so all bets are off
2026-04-28 21:45:23 <veltas      > This is one of the issues with SVG, a lot of what you'd want to do with text is create text boxes, but you can't reliably fill text boxes
2026-04-28 21:45:58 <tpbsd       > how best to solve it ?
2026-04-28 21:46:20 <veltas      > My Sans Serif font is DejaVu Serif
2026-04-28 21:46:27 <tpbsd       > veltas, thanks for testing it!
2026-04-28 21:46:49 <veltas      > Honestly tpbsd I'm not sure, but maybe if you can specify a specific font that most people have it will appear more consistent
2026-04-28 21:47:09 <vms14       > guys let's sell forth as rpn calculators :0
2026-04-28 21:47:35 <vms14       > i wonder about math and rpn, and why some people prefer it
2026-04-28 21:48:00 <vms14       > and wonder if i could actually sell my toy lang as a rpn calculator app xd
2026-04-28 21:48:44 <veltas      > tpbsd: If you can specify font you might try e.g. "Arial; Helvetica; Nimbus Sans L" because they are the same size in theory, and pretty much all systems have one of those
2026-04-28 21:49:12 <tpbsd       > ok, I'll give it a try, thanks
2026-04-28 21:49:19 <veltas      > I mean "Arial, Helvetica, Nimbus Sans L, Sans Serif"
2026-04-28 21:50:15 <veltas      > I've got Nimbus installed and none of the others, I think Windows usually has Arial and Macs tend to have Helvetica maybe(?)
2026-04-28 21:50:41 <veltas      > But Arial and Nimbus Sans L are designed to look like Helvetica
2026-04-28 21:51:46 <veltas      > Nimbus Sans L actually has exactly the same positioning dimensions as Helvetica because it's designed to be a substitute when a PDF / printed document requests Helvetica
2026-04-28 21:52:07 <tpbsd       > what a mess
2026-04-28 22:00:52 <veltas      > Maybe print your SVG in the browser to PDF...
2026-04-28 22:01:02 <veltas      > Or generate PDF in first place if that's an option?
2026-04-28 22:01:45 <tpbsd       > can you pls check it again, I've changed the font: https://raw.githubusercontent.com/techman00172/schematics/refs/heads/main/moped-batt-chgr.svg
2026-04-28 22:01:59 <tpbsd       > thats if youre not at work and have the time
2026-04-28 22:02:33 <veltas      > I'm home dude
2026-04-28 22:02:41 <veltas      > I'm not at work all day :P
2026-04-28 22:02:45 <tpbsd       > ot was monospace, Ive changed it to sans-serif9 like the rest of the text
2026-04-28 22:02:51 <tpbsd       > bargain!
2026-04-28 22:02:55 <veltas      > It's "Sans Serif" still
2026-04-28 22:03:10 <veltas      > Same exact result as before, overlapping for me because DejaVu Sans is fat
2026-04-28 22:03:27 <tpbsd       > ohh, I picked the wrong option~
2026-04-28 22:04:56 <tpbsd       > I thought that the SVG was the neatest option for people
2026-04-28 22:06:13 <veltas      > PDF is designed for this job, of displaying a document portably so it appears the same anywhere online
2026-04-28 22:06:28 <veltas      > I don't know WTF SVG is for but apparently not that
2026-04-28 22:06:33 <veltas      > Maybe vms14 can explain
2026-04-28 22:07:44 <tpbsd       > good idea!
2026-04-28 22:08:14 <veltas      > Like the reason I know about Nimbus is because I was learning about how PDF works when I looked into writing PDF generation in Forth
2026-04-28 22:09:09 <veltas      > Helvetica is one of the built-in fonts for PDF, but nobody actually really has it on their PC.  Adobe did at least specify the dimensions so that people could make look-alike fonts that sit in the same place when rendered.
2026-04-28 22:10:41 <veltas      > PDF is pretty well designed, it's not a bad format, although the original encoding is dreadful because it was designed before UTF-8 was a thing
2026-04-28 22:16:20 <tpbsd       > pdf solved the problem!
2026-04-28 22:16:32 <tpbsd       > great idea, it's more compact also
2026-04-28 22:16:49 <veltas      > Oh yeah PDF is well designed
2026-04-28 22:33:05 <veltas      > Can't send to termbin.com right now, is it working for anyone else?
2026-04-28 22:37:44 <veltas      > Fuckin pastebin doesn't even work, is the internet just being hammered right now?
2026-04-28 22:37:47 <veltas      > Bloody bots
2026-04-28 22:37:56 <tpbsd       > veltas, the war ?
2026-04-28 22:38:07 <tpbsd       > cables cut, nukes in space ?
2026-04-28 22:39:30 <veltas      > https://paste.debian.net/plainh/4f9da39c
2026-04-28 22:40:04 <veltas      > Don't scare me tpbsd
2026-04-28 22:40:27 <veltas      > I've had so many negative premonitions recently, it's been off the charts, like the sky's going to fall
2026-04-28 22:41:39 <tpbsd       > veltas, it's a scary time, the world is at war
2026-04-28 22:42:01 <tpbsd       > best be as prepared as you can
2026-04-28 22:42:14 <tpbsd       > at least mentally, if nothing else
2026-04-28 22:51:40 <veltas      > Anyway so I managed to get my generic ending syntax to work https://paste.debian.net/plainh/4f9da39c
2026-04-28 22:52:13 <veltas      > So I can write e.g.  STRUCTURE SHAPE  FIELD: WIDTH  FIELD: HEIGHT  END
2026-04-28 22:52:30 <veltas      > And use END after any syntax, not just STRUCTURE
2026-04-28 22:57:11 <tpbsd       > veltas, would you mind checking this one, how does it render ? https://github.com/techman00172/schematics/blob/main/moped-battery-charger.pdf
2026-04-28 22:58:06 <tpbsd       > sadly the easy PDF making in qucs uses inkscape and github cant render it!
2026-04-28 22:58:29 <veltas      > Looks perfect
2026-04-28 22:58:43 <tpbsd       > the harder pdf option storage location isnt non volatile!
2026-04-28 22:58:49 <tpbsd       > excellent!
2026-04-28 22:58:57 <veltas      > Better than the SVG, nice grid border too
2026-04-28 23:00:31 <tpbsd       > glad you like it!
2026-04-28 23:21:58 <veltas      > Night
2026-04-28 23:22:46 <tpbsd       > veltas, night and thanks again for your help!