Imagining the Ideal Language for Writing free Software (my talk from 2021) just got posted: bofh.nikhef.nl/events/FOSDEM/2

In that presentation, I ague that , , and aren't great fits for Free Software (even though they're all well-designed languages and I still <3 rust!). And I think about what we _do_ want in a language aiming at Free Software.

Slides are available on my site (but truly are a visual aid, so may not make that much sense alone) codesections.com/presentations

Follow

Oh, I guess that video.fosdem.org/2021/D.perl/p might be the better link for my _Imagining the Ideal Language for Writing Free Software_ talk.

I guess must be doing something fancy with redirects and load balancing, so we might as well play along.

There's also an mp4 version, if you happen to prefer that to webm: video.fosdem.org/2021/D.perl/p

@codesections thank you, really nice talk!

Its the first time I heard software development likened to authoring as a technical point, even though it has been a primary philosophical point in free/libre philosophy all along.

How complex is Raku's implementation? That aspect is one big reason I can't see sustainability in LLVM languages like Rust, or bespoke languages like Vala. Something "implementable" like Squeak Smalltalk or an RPython language seems more sustainable.

@wyatwerp OpenSmalltalk, the optimising JIT runtime usually used for Smalltalk is a few million lines of C, but I guess you can just interpret bytecode, and some day we'll see a metacircular Smalltalk system. The actual Squeak, Pharo et al images are much smaller.

@noslebadlorah i thought the opensmalltalk C code is all generated code, from slang code. Not so?

@wyatwerp github.com/OpenSmalltalk/opens has a metric fuckton of C, but you could still be right and they just condensed the Slang code down :)

@noslebadlorah they mention Github has generated code, all of it for all configurations!
--
This repository contains the code generated by the Simulator, and the platform support code for the entire VM, its CI infrastructure and so on. The core VM should not be developed by editing the generated code. The core VM should be developed using Smalltalk. The source code repository for the VM is
http://source.squeak.org/VMMaker.html
--
I didn't find any Slang or platform-specific code metrics.

@wyatwerp

> thank you, really nice talk! Its the first time I heard software development likened to authoring as a technical point, even though it has been a primary philosophical point in free/libre philosophy all along.

Thanks, I'm really glad you liked it :D

> How complex is Raku's implementation?

Medium-ish? I certainly wouldn't put in anywhere close to Squeak. But a small team has implemented it 3–5 times so far (depending on how you count, and the focus now is on ~1).

@wyatwerp

One interesting wrinkle for multiple implementations is that Raku has a spectest (github.com/Raku/roast) rather than a written spec. The pitch is that this makes multiple implementations dramatically easier (and reduces debates about what the spec "really means").

The counter-argument is that it means Raku doesn't *have* a real/traditional specification

@codesections ok. Tangentially, do you have an opinion on https://xlr.sourceforge.io ? Its FOSDEM talk is another one I want to get to listening.
Sign in to participate in the conversation
Fosstodon

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.