02:53 <Zarutian_HTC> saw recently a Forth that had 'strong typing', called play something, anyone here know anything about that?

04:50 <crc> https://www.play-lang.dev ?

04:59 <Zarutian_HTC> yeah

05:06 <Zarutian_HTC> and it is written in elm, which looks like some function programming language

05:07 <crc> Yes, written in elm, compiled to JavaScript, generates webassembly

09:08 <veltas> Real risk of it turning out the standard library is impossible to implement with strong typing :P

12:07 <veltas> "It's more important to conform, than to do it the best way you can" -- this is a statement I agree with sometimes, and sometimes I don't. It depends on context.

12:07 <veltas> I think sometimes though it does feel like I am more on the "do it the best way you can" side than most people

12:09 <veltas> Like the decision about whether to use C or one of the many other able embedded programming languages, do you conform to C or use something better?

12:10 <hosewiejacke> I don't think there are that many other embedded programming languages ;)

12:11 <veltas> There are a lot actually, that can sometimes be better than C, it depends

12:12 <hosewiejacke> Would you mind giving an example? Asm, C and Forth are the only ones I can think of.

12:15 <veltas> What's the c-killer everyone's always banging on about on hackernews?

12:15 <veltas> It's not nim or rust I think... or is it?

12:16 <veltas> Forth is an alternative to C, which you mentioned

12:16 <veltas> People use embedded versions of Python and Lua

12:17 <hosewiejacke> Oh yeah. Totaly forgot about these. I'm not that much on HN anymore.

12:19 <veltas> hosewiejacke: zig

12:19 <veltas> Look up zig I like it, not that I've used it

12:20 <veltas> The only alternative language I've given any attention is Forth, and it's because Forth is interesting even if you don't get to use it for anything practical

12:20 <veltas> And I don't think I can convince anyone at work to use any alternative to C

12:44 <veltas> From nim website: "Modern concepts like zero-overhead iterators and compile-time evaluation of user-defined functions"

17:21 <siraben> hosewiejacke: there's Zig

17:21 <siraben> veltas: ^

17:22 <siraben> oops. someone already replied :p

17:23 <siraben> C is very stable, has been proven in production, has a lot of tooling at this point for everything from memory safety to model/SMT checking, separation logic, static analysis and so on

17:23 <siraben> Would be a lot of work to bring that to Zig

17:24 <siraben> Even a formally verified C compiler exists, heh https://compcert.org/

17:50 <hosewiejacke> siraben: thanks. Yes, C has a lot of tooling; but no namespaces :p

18:17 <veltas> I work in a role where the code matters and does a crucial job and we don't have any standards for verifying stuff, we use C out of general paranoia. It's well-founded paranoia to use something that's at least conventional and bullet proof in production, but it's not exactly as formal in all of embedded as your post implies

18:17 <veltas> THere are lots of people in embedded who have flexibility to try this stuff out, the question is more about whether managers want to put their reputations/&c. on the line

18:17 <veltas> And I respect e.g. my company's use of C, it makes perfect sense and C is completely fine.

18:20 <hosewiejacke> I once asked a guy from automotive field what PL they use and he said (something along the lines): C, because it has to be compilable 20 years from now.

18:21 <hosewiejacke> I wonder if that would be the case for Rust or Zig...

18:22 <veltas> Before I worked in industry I assumed that everyone doing embedded on e.g. automation grade computers uses misra etc, I now know that very few people use stuff like that, mostly where it 'matters'.

18:23 <veltas> Also I realise Misra has a bunch of dumb rules that probably makes it harder to write safe code, if you want safe code your process needs to be smarter than "use misra" or "use cert"

21:50 <MrMobius> there was a funny thing where they looked at the rate of bugs in one thing then looked at what happened in misra projects and it looked like it didnt actually eliminate many bugs

21:50 <MrMobius> but maybe the two measures arent exactly equivalent