IRC Log - 2025-08-05 - ##forth

Channel: ##forth
Total messages: 182
Time range: 00:01:28 - 23:58:29
Most active: tpbsd (102), veltas (44), xentrac (27)
00:01:28 ##forth <veltas> So you can get mod 5 by e.g. doing 16-bit sum of the bytes of the number; then doing an 8-bit sum of the nybbles etc or however you want to skin the cat
00:02:00 ##forth <veltas> I'm sure there's lots of info on how to do this but it was just a little interesting puzzle for me, not really doing anything productive
00:03:22 ##forth <veltas> Getting the mod on its own isn't helpful, but I wasn't sure whether you could get the mod 'faster' somehow than dividing, and yes indeed you can get mod a lot faster
00:07:01 ##forth <veltas> The mod result I googled about after and read about Chinese Remainder Theorem, which I think I 'get' a bit better after the exercise
00:40:17 ##forth <tpnix> xentrac, the Mecrisp-Stellaris LangServer is just to test things, the main project also includes all the registers and bitfields of a Cortex-M mcu by using the CMSIS-SVD file
00:41:04 ##forth <tpnix> xentrac, so a Forth embedded dev will now have completion on everything, the Mecrisp-Stellaris Forth commands and the mcu hardware
00:41:49 ##forth <tpnix> along with examples in the LSP popup
09:04:30 ##forth <veltas> Released v0.1.0 of zenv https://github.com/Veltas/zenv/releases/tag/v0.1.0
09:16:52 ##forth <tpbsd> veltas, good job!
09:19:56 ##forth <tpbsd> Forth for the zx spectrum!
09:20:59 ##forth <tpbsd> this relates to the Z80 code you mentioned earlier
10:00:28 ##forth <veltas> Actually unrelated but related in the sense it's for the same platform
10:00:56 ##forth <veltas> I mostly did the release so I could download the binaries on my phone easier, I've not done a tagged release before
10:01:22 ##forth <veltas> I fixed a bug in number input as an excuse
10:20:46 ##forth <veltas> And thanks!
10:22:30 ##forth <tpbsd> it's great to see people developing Forth projects :)
10:23:10 ##forth <tpbsd> because it's projects that make Forth more visible in general
10:23:27 ##forth <tpbsd> even tho Forth is in the main, generally avoided thesedays
10:25:08 ##forth <tpbsd> I played a bit of a joke on all the arduino people a few years ago, I wrote the "bluepill diagnostics" in Forth and released it, SF gave me a 10,000 downloads badge a couple of years ago and it's had a steadt 6ish downloads a day since then
10:25:09 ##forth <veltas> Technically you could use that Forth to do embedded projects, although I can't think of a good reason to rely on a ZX Spectrum for that
10:25:28 ##forth <veltas> Wow
10:26:37 ##forth <tpbsd> my program interrogates a 'blue pill' board $2 -5 to find what mcu is actually in it as unscrupulous Chinese vendors were fitting fake renamed mcus instead of STM32F103's
10:27:17 ##forth <tpbsd> and lots of the bluepills werent talking on the USB so there was much confusion
10:27:51 ##forth <tpbsd> it was a piece of cake to use Mecrisp-Stellaris to do that job as embedded Forth is just so easy
10:29:30 ##forth <tpbsd> I coded a simple but unbreakable menu screen for the terminal and made sure to announce this was 'powered by Forth' in it
10:29:41 ##forth <veltas> Lovely
10:30:30 ##forth <tpbsd> you can see the menues in that url user manual
10:31:04 ##forth <tpbsd> it's the defacto method of testing bluepills now as it works everywhere
10:31:40 ##forth <tpbsd> they just plug in a usb cable and run a term on the pc, windows,mac or Linux doesnt matter, it's the same
10:31:57 ##forth <tpbsd> up pops the menu and then they run the tests
10:36:52 ##forth <veltas> I think when you mentioned this years ago that was one of the first times I learned about Chinese "good enough" manufacturing
10:37:30 ##forth <veltas> Seems like it's better understood now in the mainstream, big subject of trading controversies with China right now
10:39:00 ##forth <tpbsd> actually it was just a few lowlifes doing it and even back then the Chinese authorities were trying to stamp them out
10:39:27 ##forth <tpbsd> companies like GD were actively pursuing the culprits
10:39:50 ##forth <veltas> I'm hearing that it's a nightmare getting chinese manufacturers to not do this in general, but I don't doubt their authorities oppose it
10:40:00 ##forth <veltas> Because it's not really good for the country as a whole, if nobody trusts them
10:40:24 ##forth <tpbsd> and in the end, the Chinese STM32F103 clones are licensed, way cheaper, faster and they solved all the silicon problem of the original
10:40:28 ##forth <tpbsd> true
10:40:35 ##forth <tpbsd> they di a lot of damage
10:40:36 ##forth <veltas> It seems like a cultural thing rather than a state-approved practice
10:40:56 ##forth <veltas> It only takes a small % of untrustworthy people to lower trust generally
10:41:05 ##forth <tpbsd> true
10:42:20 ##forth <tpbsd> but I eventually overcame my bias, I only buy from china now
10:43:20 ##forth <tpbsd> I bought some laser distance measurers from china, theyre the size of a usb stick, perfectly made with a small lcd with high res to display the data
10:43:46 ##forth <tpbsd> they do 30 metres at 1mm accuracy, and cost $27 AUD each
10:44:05 ##forth <tpbsd> the level of tech is amazing, quality is 100%
10:44:15 ##forth <tpbsd> theyre usb and rechargable
10:48:54 ##forth <veltas> It's a simple equation for me, most stuff is cheaper in China so I'll buy most stuff from them, as long as the cost of quality issues doesn't get in way
10:49:05 ##forth <veltas> But so far it's not caused too much trouble
10:50:08 ##forth <tpbsd> I always buy the best I can afford at the time, it's how I was brought up
10:50:21 ##forth <tpbsd> I never buy the cheapest, thats my last option
10:51:00 ##forth <veltas> Well I buy cheapest overall cost I can get away with
10:51:04 ##forth <tpbsd> but China really throw a spanner in the works because they have asesome tech for cheap and junk for cheap as well
10:51:29 ##forth <veltas> E.g. I bought my shoes for like 350 GBP on basis that they would be cheaper in long run, and I think that logic has broken even already
10:51:38 ##forth <veltas> So hopefully they continue to last
10:52:06 ##forth <veltas> Had a new heel and resoled once and still cheaper than buying the crap that was falling apart in less than a year
10:53:54 ##forth <tpbsd> wat to go!
10:53:58 ##forth <tpbsd> way
10:56:49 ##forth <veltas> If people don't like us buying from China I would say we need to figure out how we can compete with them, rather than trying to manipulate prices to make people avoid them
10:57:12 ##forth <veltas> That's a deeper question, why can't we do what they do. We need to solve that question if we want to dominate the market in future.
11:03:59 ##forth <tpbsd> veltas, absolutely
11:04:57 ##forth <tpbsd> if you want people to buy your product, fair price and high quality will mean you dont need to advertize
11:05:58 ##forth <tpbsd> when I wanted a pushbike in 2000, I paid $3500 for a Cannondale 'made in the USA' and that bike is still immaculate 25 years later, and the places it's been!
11:06:32 ##forth <tpbsd> Cannondale had a rep for high quality, their people took pride in their work
11:06:58 ##forth <tpbsd> canondale is now made in china
11:07:37 ##forth <tpbsd> the bikes look great, theyre really high tech, Id buy one if I had a spare $14,000 for the electric mtb
11:07:53 ##forth <tpbsd> they do 100 miles on a charge they claim
11:10:47 ##forth <veltas> Nice
11:11:45 ##forth <veltas> Yeah reputation unfortunately only matters as much as your consumers are informed, and people are just a lot less informed. But it still has an impact.
11:11:48 ##forth <tpbsd> theyre awesome imho
11:12:13 ##forth <tpbsd> reputation drove sales before the Internet
11:12:22 ##forth <veltas> Looking at state of the car industry etc, the people buying are often more interested in brand names than quality etc
11:12:23 ##forth <tpbsd> you cant stop good or bad news
11:12:43 ##forth <tpbsd> thats probablly advertising ?
11:13:10 ##forth <veltas> I think brands are probably a result of quality over time, and then when the brand is 'good' they slip
11:13:30 ##forth <veltas> Not always but I would say so with e.g. Audi
11:13:46 ##forth <veltas> Seem to be very pricey and the product isn't that good compared to similarly priced offerings
11:14:25 ##forth <tpbsd> veltas, I have a Audi :)
11:15:31 ##forth <tpbsd> veltas, it was $60,000 AUD in 2008, it's permanent 4wd, goes like a rocket and is the best handling car I've ever driven
11:15:44 ##forth <tpbsd> it's a 2.8l V6 petrol
11:16:11 ##forth <tpbsd> Ive driven it down the beach, it doesnt even leave tracks
11:16:52 ##forth <veltas> It seems like the quality was good in 2008
11:17:04 ##forth <veltas> But I'm just hearing it's really overpriced now, for what it is
11:17:15 ##forth <tpbsd> each wheel hd souble wishbone suspension, and it has 'torsen' diffs. That means if you lift one wheel, it will still rotate under power
11:17:33 ##forth <tpbsd> yeah, all cards are crap now, inc audi
11:17:37 ##forth <tpbsd> cars
11:17:47 ##forth <veltas> It's inevitable really in business that without proper leadership they'll just cannibalise their brand and reputation
11:17:54 ##forth <tpbsd> true
11:18:14 ##forth <tpbsd> busines in the west is disintegrating fast sadly
11:18:27 ##forth <tpbsd> while it's going ganbuster in the east
11:18:42 ##forth <veltas> Yeah and it's a bit scary, we need to get a grip really
11:19:03 ##forth <veltas> I don't personally think the solution is protectionism
11:19:07 ##forth <tpbsd> tho one western biz is way ahead ... spaceX
11:19:17 ##forth <tpbsd> I agree, thats not the answer
11:20:04 ##forth <tpbsd> I use Starlink and it's awesome for me as in in a remote rural ausie town with no copper or fibre and only 'wireless' Internet
11:20:18 ##forth <veltas> Nice
11:20:27 ##forth <tpbsd> Ive downloaded two 60 GB I images just today
11:20:34 ##forth <tpbsd> AI images
11:20:49 ##forth <tpbsd> Musk is a trebdsetter for sure
11:21:18 ##forth <tpbsd> Starlink now has 6 million users worldwide and 5000 sats in orbit
11:21:38 ##forth <tpbsd> thats 6 million times 130AUD = $LOTS
11:21:42 ##forth <tpbsd> a month
11:21:54 ##forth <veltas> He's a very proficient business person, and good at running engineering companies. But he also is a total wildcard and odd. He needs people around him to get him to focus on positive things and not self-destructive things
11:22:11 ##forth <veltas> Like a lot of eccentric brilliant people
11:22:15 ##forth <tpbsd> yeah, but many like him ar like him
11:22:27 ##forth <tpbsd> look at howard hughes ?
11:23:12 ##forth <tpbsd> tho Robert Noyce was a gift to the USA as he started Fairchild, Signetics, Intel etc
11:23:20 ##forth <tpbsd> and other probably
11:24:01 ##forth <veltas> It's so easy to be critical, we should learn from the bad and be inspired by the good
11:24:44 ##forth <tpbsd> veltas, heh, so in my own way, I got 10,000+ Forths into the hands of mostly Windows Arduino users over the last 4 years, even if they didnt know it
11:24:59 ##forth <tpbsd> agreed
11:25:21 ##forth <tpbsd> you and I have very similar philosophies I think
11:25:56 ##forth <veltas> History says that means we'll have the most prejudice against each other!
11:26:06 ##forth <tpbsd> I think Musk could all by himself get America productive again if he had the chance
11:26:35 ##forth <tpbsd> just dont judge him for being a weird, self centred, rich nastard
11:26:41 ##forth <tpbsd> bastard
11:27:41 ##forth <tpbsd> sadly the USA infrastructure is all old and run down. there is no way to 'get America working' untill the supply chain is fixed
11:28:00 ##forth <tpbsd> America really needs to copy china
11:28:58 ##forth <tpbsd> you should sack trump amd hire Jack Ma, pay him what ever he wants to fix the place up
11:29:11 ##forth <tpbsd> Jack Ma set up Alibaba etc
11:38:28 ##forth <tpbsd> veltas, this YouTube is exactly about a new USA firm making bycycles (they were in china) https://www.youtube.com/watch?v=QuTxTYNcIjU
11:42:58 ##forth <veltas> I'll watch after work, thanks
11:43:51 ##forth <tpbsd> no problemo
20:23:46 ##forth <xentrac> veltas: hmm, I'll have to think about that algorithm
20:24:29 ##forth <xentrac> tpbsd: autocomplete is good!
20:25:59 ##forth <tpbsd> xentrac, especially when one has 15,000 names to remember !
20:26:26 ##forth <tpbsd> xentrac, and a lot of them are not designed to be remembered
20:27:18 ##forth <xentrac> I was going to say, are you talking about things like the GD32F103, which is a faster clone of the STM32? but no, you mean people who mislabeled their clone chips as ST chips
20:28:24 ##forth <tpbsd> yes, I mean people who deliberately removed the GD32F103 markings and printed STM32F103 markings to sell more product
20:29:28 ##forth <tpbsd> heh, all the STM32F103 clones are faster and cheaper, and have more memory
20:30:20 ##forth <xentrac> really? I didn't know that
20:32:24 ##forth <tpbsd> yeah, the most compatible choice is the APM32F103
20:32:42 ##forth <tpbsd> you can get them from lcsc.com
20:32:54 ##forth <xentrac> I think maybe you mentioned the APM32 the other day and it was new to me
20:33:09 ##forth <tpbsd> dont use the GD32F103 as it has SPI flash and isnt fully compatible
20:34:03 ##forth <tpbsd> the APM32F103 is in all the Chinese "thumbdrive like' STLINK units in the pretty anodised aluminium cases that sell for $2
20:34:05 ##forth <xentrac> I've heard that. does it take a long time to boot? I understand it's faster when it's running
20:34:29 ##forth <xentrac> (the GD32)
20:34:36 ##forth <tpbsd> true for all mcus that use a SPI flash as rom
20:34:42 ##forth <tpbsd> including the Pico
20:34:49 ##forth <xentrac> presumably when you power it up it copies the Flash into RAM?
20:34:53 ##forth <tpbsd> slow as in 500 ms
20:34:59 ##forth <tpbsd> exactly!
20:35:37 ##forth <tpbsd> the APM32F103 has internal flash so it's boot up = that of the STM32F103
20:36:01 ##forth <tpbsd> which is probably 0.5uS
20:36:17 ##forth <xentrac> 500ms is a long time especially if you're shooting for an 0.1% duty cycle for power consumption reasons
20:40:40 ##forth <tpbsd> yeah, the fans never seem to know that which means theyre hobbyists with no real embedded hardware experience or programmers only
20:42:04 ##forth <tpbsd> of course, your cat can wait 500mS for it's automatic door opener to activate, Im sure it wont mind ;-)
20:42:51 ##forth <xentrac> 500 millisiemens is a pretty low resistance path
20:43:01 ##forth <tpbsd> oops
20:43:07 ##forth <xentrac> 500mʊ
20:57:39 ##forth <tpbsd> xentrac, as a developer you may be interested to see the my just finished article on how one can make kickarse readme.md's with an ai ? https://discourse.homelabbrisbane.com.au/t/an-ai-documentation-masterclass-for-humans/226/1
20:58:01 ##forth <tpbsd> frankly, Im stunned
20:59:02 ##forth <tpbsd> haha, programmers should be embracing AI, ... for user doc creation (which they hate and are terrible at anyway)
22:14:31 ##forth <veltas> Watched the USA bike factory video, was interesting
22:21:28 ##forth <xentrac> tpbsd: I like user doc creation
22:22:04 ##forth <xentrac> but I've definitely found LLMs useful for fleshing out docs and for writing example code
22:22:50 ##forth <xentrac> I can't really read your fixed-format blockquotes because they're too wide for my screen
22:24:14 ##forth <xentrac> I mean once Discourse gets done filling up the left fourth of the screen with a navigation sidebar, another twelfth with a column for your usericon, and another eighth on the right side for a redundant scrollbar, there isn't enough space left over on my screen to fit what looks like 80-column text
22:24:51 ##forth <xentrac> if I grab the scrollbars on the fixed text areas I can drag them left and right once per line in order to read the text one line at a time
22:25:44 ##forth <xentrac> except that some of them are so tall that by the time I scroll down to the horizontal scrollbar I can no longer see the top of the text area
22:25:47 ##forth <xentrac> such as "A fresh, user-friendly `readme.md` file has been successfully written to `/home/tp/fossil/mecrisp-stellaris-lsp/readme.md`, replacing the previous version."
22:26:34 ##forth <xentrac> so in theory I could read that line by reading the left half of it, scrolling down to the horizontal scrollbar, scrolling the scrollbar all the way to the right, scrolling back up to the line to read the right half of it, and then repeating the process for each following line
22:28:02 ##forth <xentrac> although now that I have pasted the line, it looks like that line is too wide to fit on my screen even without wasting half of the width of my screen on window furniture
22:28:36 ##forth <xentrac> I mean it's almost two full lines in this IRC text window
22:28:58 ##forth <xentrac> I'm curious what kind of design process at Discord led to this user experience trainwreck
22:33:45 ##forth <tpbsd> xentrac, that server belongs to a friend, yeah the Discord server is pretty low quality
22:34:40 ##forth <tpbsd> xentrac, it's even stopped accepting image uploades so I suppose it will crash soon
22:38:56 ##forth <xentrac> I don't think Discord servers are actually servers; they're just like virtual hosts in the Discord cloud
22:39:14 ##forth <xentrac> like, you can't download a copy of the Discord software to install on your own computer (except the client)
22:40:06 ##forth <xentrac> you could probably ameliorate the problem by not using fixed-width text for blockquotes, a thing for which I think Discord has its own markup
22:49:08 ##forth <tpbsd> ahh, good idea
22:49:29 ##forth <tpbsd> oh I was very ignorant about discord, thanks
22:49:54 ##forth <xentrac> I'm no expert, but I think that
22:49:57 ##forth <xentrac> s how it works
23:55:42 ##forth <cleobuline> forthBot: S" un monstre affreux avec des yeux globuleux des grandes dents et des tentacules "S IMAGE
23:56:54 ##forth <forthBot> DEBUG: No 'url' key in ImgBB response
23:56:55 ##forth <forthBot> Error: IMAGE: Failed to generate or upload image
23:57:15 ##forth <cleobuline> forthBot: S" un monstre affreux avec des yeux globuleux des grandes dents et des tentacules "S IMAGE
23:58:28 ##forth <forthBot> DEBUG: No 'url' key in ImgBB response
23:58:29 ##forth <forthBot> Error: IMAGE: Failed to generate or upload image