2026-02-06 03:40:29 <tpbsd       > Aaron Viejo is developing mecrisp-ice on Lattice Certus using  mecrisp-ice-master/hx8k for anyone interested
2026-02-06 03:41:06 <tpbsd       > https://sourceforge.net/p/mecrisp/discussion/general/thread/bb53a98cf3/?limit=25#5b4b
2026-02-06 05:07:32 <iv4nshm4k0v > All the Office Open XML formats represent their data as files in a .zip file.  So are OpenDocument ones, as well as .epub, and likely a number of other formats I can't recall off the top of my head...  Right, just checked http://en.wikipedia.org/wiki/ZIP_(file_format) : Java .jar and Android .apk use .zip as well.
2026-02-06 10:09:53 <veltas      > iv4nshm4k0v: I think my main point is that MS hasn't provided a better way to do this than something that's for sure very easy for the likes of you and me, but isn't discoverable or relevant to 99% of their intended audience
2026-02-06 10:10:22 <veltas      > But it was a nice twist about the .zip thing; I think I'd heard this before somewhere but I'd never put two-and-two together
2026-02-06 10:11:08 <veltas      > Although .zip isn't an amazing archive format but it's better than .tar.gz, provided you ban certain unsafe 'features'
2026-02-06 10:12:02 <veltas      > (if someone wonders why I criticise .tar.gz: I love simplicity of .tar.gz but it's much better at whole compression/extraction rather than individual files)
2026-02-06 10:25:45 <KipIngram   > You mean that .tar.gz is good for "whole archives" but hard to treat like a "file repository"?  I don't know that I have ever tried to manipulate a single file in a .tar.gz.
2026-02-06 10:53:14 <veltas      > Correct
2026-02-06 10:53:37 <veltas      > You can do it, tar makes it easy to do this, but it's not as performant as a .zip which is better at 'walking' through files
2026-02-06 10:53:58 <veltas      > To 'walk' through .tar.gz as far as I know you must decompress the entire stream until you strike gold
2026-02-06 11:57:21 <iv4nshm4k0v > I frankly find it difficult to be surprised at what MS does as I frankly find it difficult to understand what they're /trying/ to do.  Though I find office productivity suites rather unproductive in general, regardless of who produces them.
2026-02-06 11:57:22 <iv4nshm4k0v > As to .zip archives, I find them good for what they do: allow transferring a set of files with /some/ compression, interoperably.  The format is supported on pretty much every general purpose computer starting with 16-bit ones.  AIUI, 8-bit machines (say, those running CP/M) require different formats.  (Though note that ISO/IEC 21320-1 used in ODF et al. is a subset of the original APPNOTE.)
2026-02-06 12:06:26 <veltas      > I'm biased as a programmer but I prefer documents to be based on some source text/program, so I can apply the same tools to manage them I do with general purpose programs
2026-02-06 12:06:54 <veltas      > And so I can describe what I want rather than constantly wrestle the WYSIWYG
2026-02-06 12:08:41 <veltas      > A middle ground is something like Adobe FrameMaker
2026-02-06 12:09:33 <veltas      > Where it's technically probably possible to manage the XML with a version control system, but it's still WYSIWYG combined with a structured tree of content, you've got control but you don't have to be a programmer to figure it all out
2026-02-06 12:09:38 <veltas      > In theory
2026-02-06 12:10:17 <veltas      > The problem is I think often a really simple programming library is easier even for non-techies to understand than a massive DSL or graphical program
2026-02-06 12:11:01 <veltas      > Because it turns out we all need the same flexibility and the limit should be our imagination, not the number of 'english-like' things we can memorise or dark corners of a GUI we need to explore
2026-02-06 12:11:31 <veltas      > These DSLs quickly devolve into bad programming languages
2026-02-06 13:32:39 <iv4nshm4k0v > 
ACTION listens to http://opengameart.org/sites/default/files/inferior_diety_v2.ogg 

2026-02-06 13:32:39 <iv4nshm4k0v > My overall preference is to use GUIs strictly for things inherently graphical.  And even then, I'd argue some common GUI conventions should be used only sparingly, if at all, such as icons and multi-level menus.  Darktable offers a fairly good GUI, IMO - e. g., one with a text filter for "plugins" rather than a plain old menu; Kicad, much less so.
2026-02-06 13:35:02 <iv4nshm4k0v > As to DSL, well, I'm not a webapp expert, but Javascript has originated, AIUI, as a DSL for HTML - and has evolved into a fairly decent language.  At least as of http://262.ecma-international.org/5.1/ - I'm much less certain about newer versions, or how fit it is a "WSL" proper.
2026-02-06 13:38:44 <veltas      > JavaScript I wouldn't have ever called a DSL
2026-02-06 13:39:38 <veltas      > From the beginning they understood it had to be essentially a general-purpose scripting language, it's not dissimilar to writing other scripting languages of the era, with a C-style syntax
2026-02-06 13:39:46 <veltas      > I think this has been a good decision on their part
2026-02-06 13:40:03 <veltas      > I know originally they wanted to use lisp-style syntax, what a different world that would have been!
2026-02-06 13:41:39 <veltas      > There are some DSL features but not the kind of thing I'm thinking of, nobody ever claimed javascript was accessible to non-programmers, it never pretended not to be for programmers or a programming language
2026-02-06 13:42:36 <veltas      > It 99% looks like normal programming, the only big difference I can think of is that it's obviously more event-driven than a normal general programming language would be
2026-02-06 13:43:04 <veltas      > But that's similar to e.g. win32 programming which is also event-driven, but done in C
2026-02-06 13:45:43 <iv4nshm4k0v > I can't quite recall when I've first became interested in JS, but I think it was after ES 5.1 was published in 2011.  I seem to - vaguelly - recall that early JS versions were fairly idiosynchratic.
2026-02-06 13:49:56 <veltas      > There's only a few weird things caused by implicit type conversions, overly overloaded operators, and var being function-scope regardless of the block it appears in
2026-02-06 13:50:20 <veltas      > Because it was made in like a month by one person, so it's pretty good considering
2026-02-06 13:51:12 <veltas      > It repeats a lot of the same confusing nonsense that scripting languages do to support 'classes' like prototypes etc
2026-02-06 13:51:19 <veltas      > What can you do eh
2026-02-06 13:54:04 <MrMobius    > isn't the comparison of strings to numbers weird and it has both == and ===?
2026-02-06 13:58:33 <veltas      > == I think does a lot of implicit conversions, overly aggressive sometimes
2026-02-06 13:59:03 <iv4nshm4k0v > That objects get converted to strings when needed is a fairly common feature for scripting languages (Tcl has "everything is a string" as one of its core principles; not to mention Unix shells), and == / === is similar to Lisp's EQUAL / EQ, so not something I've ever found trouble with.
2026-02-06 14:00:25 <veltas      > It's nice when you want it, it's not good if someone's flying blind and doesn't know the language well
2026-02-06 14:00:34 <veltas      > When is that not true though
2026-02-06 14:00:55 <iv4nshm4k0v > FWIW, you can compare strings (i. e., their addresses) with = in Forth, but you can also compare them character-to-character.
2026-02-06 14:03:12 <veltas      > Someone who doesn't know what they're doing is far more dangerous in Forth
2026-02-06 14:03:36 <veltas      > Or maybe less dangerous because they'll hit a brick wall of segfaults faster(?)
2026-02-06 14:04:37 <veltas      > Not unreasonable to expect people to learn a language before using it, the unreasonable bit is how many languages you need to learn to do even basic things today :)
2026-02-06 14:06:04 <KipIngram   > Heh.  Right - the most dangerous may be the person who knows "some."
2026-02-06 14:06:13 <KipIngram   > Enough to cobble something working together.
2026-02-06 14:06:43 <KipIngram   > So, I have decided that the most important purpose of the right arrow key on my notebook is fast forwarding through sponsor sections of YouTube videos.
2026-02-06 14:07:11 <KipIngram   > UBlock Origin just annihilates "actual commercials" - as soon as I installed it years ago they simply disappeared from my experience.
2026-02-06 14:07:36 <KipIngram   > But that doesn't catch the ones where the video presenter talks about the sponsor's product themselves.  Fast forward is required for that.
2026-02-06 14:07:51 <veltas      > I think I use L for that(?)
2026-02-06 14:08:02 <KipIngram   > There's an AI tool we need - one that can recognize those sections and automatically jump over them.
2026-02-06 14:08:03 <veltas      > It's all muscle memory, I don't even know
2026-02-06 14:08:08 <KipIngram   > Right.
2026-02-06 14:08:10 <veltas      > Yeah I hate sponsored segments
2026-02-06 14:08:13 <veltas      > You're not wrong
2026-02-06 14:08:20 <KipIngram   > I didn't really l could do it.  I'll have to check.
2026-02-06 14:08:43 <KipIngram   > But in this case my "home typing position" fixation doesn't really apply, since I'm not generally "on the keyboard" at the time anyway.
2026-02-06 14:09:18 <KipIngram   > Most annoying is when I'm listening with ear buds and have walked away.  Usually I'll tap the bud to pause until I get back if a sponsor section starts.
2026-02-06 14:09:41 <KipIngram   > It's all so phony - you  KNOW that all of these people who talk about Brilliant aren't actual Brilliant users.
2026-02-06 14:09:49 <KipIngram   > But they pretend they are.
2026-02-06 14:09:51 <veltas      > Looks like L moves 10 seconds at a time, and right-arrow is 5
2026-02-06 14:10:07 <veltas      > So maybe now L will be your most useful key
2026-02-06 14:10:13 <veltas      > K pauses and J rewinds
2026-02-06 14:10:16 <KipIngram   > Might be.  Thanks.
2026-02-06 14:11:57 <veltas      > K is better for pause because Space randomly does other things depending on focus, the worst K can do is type 'k'
2026-02-06 14:12:16 <veltas      > I don't use J/L religiously
2026-02-06 14:13:08 <veltas      > Sponsor sections are such rubbish, even if they use the products they wouldn't have used them if they weren't being paid to
2026-02-06 14:13:32 <veltas      > It's insulting to the viewer, I want my money back!
2026-02-06 14:14:24 <veltas      > I don't really mind them because people make money from them, and I'm not paying to watch.  I leave ads on for people to make a penny or whatever.  But I don't really care if others do the same, and I'd never enable ads or do a sponsorship if I had a channel
2026-02-06 14:14:33 <veltas      > Because I don't think it's good content
2026-02-06 14:15:53 <veltas      > And I'd never buy ads from google etc, they clearly are bad value, I get shown the same ads continuously on my TV and I now hate all 2 products being shown
2026-02-06 14:35:59 <iv4nshm4k0v > That reminds me that I've been using mpg123(1) for 20+ years and never needed any of its control keys.  (In fact, my scripts would mostly invoke it like 'mpg123 -- - < "$file" ' anyway.)  Then I've had to learn "fast forward" when I've started listening to http://archive.org/details/OTRR_Dragnet_Singles .
2026-02-06 14:51:38 <KipIngram   > Hmmm.  I think I need to add RISCV to my "targets."  The ESP32-C6 has 512k RAM, several hundred kB of ROM, and sports WiFi and Bluetooth ready to go out of the box.  About the size of a coin - $6.
2026-02-06 14:51:52 <KipIngram   > That's... very nice.
2026-02-06 20:20:47 <tpbsd       > KipIngram, one problem I have with the ESP* is that theyre highly proprietary and contain a buried supervisor prpgram that steals a tiny amount of time from your application, making realtime impossible
2026-02-06 20:24:10 <vdupras     > I have a question about the word "realtime". When we speak of realtime, isn't it always "realtime within X", X being a time span? If that supervisor "steals" time in a way that is predictable, it doesn't necessarily make "realtime" impossible, but would only make "X" bigger, no?
2026-02-06 20:50:14 <veltas      > iv4nshm4k0v: Did you settle on a BSD?
2026-02-06 20:54:41 <veltas      > vdupras: In real-time systems the time constraint doesn't change, but a real-time system might take different amounts of time or be non-deterministic
2026-02-06 20:54:55 <veltas      > As long as you can prove it satisfies the timing requirement it's still real-time
2026-02-06 20:55:05 <veltas      > But if the time requirement needs to increase to satisfy the system, it's not real-time
2026-02-06 21:06:24 <vdupras     > that was also my understanding. Hence, we could say, for example, that a routine that has a binary code that has a worste case runtime of 10ns, if the supervisor can steal up to 5 ns every second, then a "realtime within 15ns" is possible for this routine on this CPU
2026-02-06 21:10:40 <KipIngram   > vdupras:  Right - if the interruptions are "relaible," then it wouldn't totally destroy real-time ability; it would just lower how "tight" your real time constraints could be.  It's very much dependent on exactly  how it works.
2026-02-06 21:11:11 <KipIngram   > tpbsd:  Does that supervisor still run even if I completely take over the processor and program in assembly?
2026-02-06 21:11:54 <tpbsd       > I read that it is completely independant and cant be accessed at all
2026-02-06 21:11:59 <KipIngram   > Or is it embedded in the libraries they offer for hgher level languages?  "Propritary" bothers me - do they not offer the necessary documentation to take direct control of everything yourself?  That's a show-stopper.
2026-02-06 21:12:15 <KipIngram   > That makes it sound like it runs on separate hardware.
2026-02-06 21:12:27 <tpbsd       > they definitely don't allow full control
2026-02-06 21:12:36 <KipIngram   > That's very unfortunate.
2026-02-06 21:12:48 <KipIngram   > Completely contrary to the whole mind-set I'm going for.
2026-02-06 21:13:09 <tpbsd       > yeah, it's a seperate subsystem that has priority control over the other processors
2026-02-06 21:14:39 <tpbsd       > Ive seen screenpics of scopes monitoring gpio outputs that 'freeze' at regular times as the 'supervisor' does stuff
2026-02-06 21:15:19 <tpbsd       > it was at that point that I totally lost interest my projects assume 100% control of the cpu
2026-02-06 21:16:03 <tpbsd       > if Im polling a switch, I wont allow something to take control of my GPIO for it's own reasons
2026-02-06 21:17:41 <tpbsd       > I think the ESP* is a amazing bit of gear for the price, very clever design just not something I'd use. Ive been given a couple of them and have never spent time 'playing' with them
2026-02-06 21:19:04 <tpbsd       > I also dont believe that wifi has any place in slow resource limited mcu's, but I think Lora etc is perfect, ie very slow speed for sensors etc
2026-02-06 21:21:05 <tpbsd       > zigbee is also ok imho