Conversation

Notices

1. Embed this notice
   ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:04:44 JST ThePhD

    Why is Fortran still used in so many places?

   Eternal question I'd actually like an answer to, so here I go reading!!

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Friday, 29-Dec-2023 14:04:41 JST Rich Felker

     in reply to

     • jade

     @leftpaddotpy @thephd The lack of LPM and language specific build system (vs utilizing whatever language agnostic programming agnostic system for declaring derivation rules you like) are some of the best things about C.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     jade (leftpaddotpy@hachyderm.io)'s status on Friday, 29-Dec-2023 14:04:42 JST jade

     in reply to

     @thephd also the terrible c packaging problems are only 80% c packaging problems; the real issue there is the crap can be built by any build system, making interop and dependency declaration hard which *actually* implies this is simply a large software problem and not a c problem

     any sufficiently complex software will have a fucked up build process

   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:04:43 JST ThePhD

     in reply to

      For example, the C & C++ ecosystems never managed to standardise any sort of build & distribution mechanism, and since Python is wrapping around basically any flavour of C & C++ code, the problems in Python packaging are a superset of what's already a hugely pressing issue in C/C++.

     God yes. Finally, somebody understands why a lot of us keep going down to the barebones levels to get things done. You can't just Rust/Java/etc. you're way out of the fundamental computing problems, buckaroo, you've got to face the music eventually!!!

   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:05:25 JST ThePhD

     in reply to

     While Fortran has long been the butt of the joke in IT departments the world over, in a curious twist of fate, it has seen a dramatic resurgence over the last few years. …

     News of Fortran's death were greatly exaggerated, etc. etc.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:06:36 JST ThePhD

     in reply to

     Same with C's death, of course. God what I wouldn't give for C to be well and truly dead, but every Professor, PL engineer, etc. all high on the supply kept trying to sell me on the fact that C (and C++) was gonna die, except it never did lmaoooo.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:06:37 JST ThePhD

     in reply to

     … LLVM, and its flagship compiler Clang. This was an absolutely massive undertaking, but due to its cross-platform nature, permissive licensing, and modular infrastructure, has become the focal point.…

     And thank god, because GCC would've been dicks if that was the standard.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:06:38 JST ThePhD

     in reply to

     As such, there was renewed vigour in the Fortran compiler space, and several important developments happened in short succession.

     YEEAH FORTRAN COMING BACK LIKE A PHOENIX FROM THE EGGY ASHES BAYBEEEEEEE

   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:07:42 JST ThePhD

     in reply to

     55'000 tests, and not a single failure? When does that ever happen on the first try after major surgery?

     Holy shit. they pulled it off. Flang did it.

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     ThePhD (thephd@pony.social)'s status on Friday, 29-Dec-2023 14:07:43 JST ThePhD

     in reply to

     LLVM provides several so-called "Intermediate Representations" (IRs), which are somewhere in between machine code and the code that most programmers write, but in a way that is language-agnostic.

     (* Rust people break in through the door *) "LIKE HELL IT IS, LLVM IS SO C-BRAINED." (The Rust people are correct.)

   • Embed this notice
     Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 29-Dec-2023 14:09:12 JST Haelwenn /элвэн/ :triskell:

     in reply to
     @thephd It's funny they mention the build & distribution mechanism of Python of all the cursed ones that never settled down for more than few years and with PEP517 decided to just explode in a bunch of tiny pieces.
     
     Source: https://xkcd.com/1987/
     Transcript: https://explainxkcd.com/1987/#Transcript
     python_environment.png
   • Embed this notice
     GeePawHill (geepawhill@mastodon.social)'s status on Friday, 29-Dec-2023 16:32:00 JST GeePawHill

     in reply to

     • Haelwenn /элвэн/ :triskell:

     @lanodan @thephd Idunno. As far as I can tell package management is still an open problem in computer science.

     Every environment I've ever worked in, the worst part of it was managing packages and dependencies.

   • Embed this notice
     Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 29-Dec-2023 16:32:00 JST Haelwenn /элвэн/ :triskell:

     in reply to

     • GeePawHill
     @GeePawHill @thephd Nah, it's a case of many language designers not looking at prior art and preferring to make their own wheels all the time.
     
     Like distros are by far the most boring choice (except on I'd say Debian where it's too painful) and the only right kind of tooling to rely on if you have system dependencies.
     
     That said I'd also point at CPAN as one that's probably the best.
     But even in the most hype ones, npm (13+ years old already) only has a lack of network-isolated builds in the reference implementation (allowed the leftpad chaos), and sadly on the non-technical side probably the most extreme case of trust issues on dependencies.