00:08 <superherointj> doublec, sent the patch. https://github.com/doublec/pony-sdl-example/pull/1

00:38 <doublec> superherointj: thanks for that, merged.

00:38 <doublec> superherointj: yep I still use it.

00:40 <superherointj> doublec, I found your blog quite interesting. You seem to have a large experience with different languages. What is your current view on Pony? I'm new to Pony. Where did it come short in your view?

00:45 <bougyman> doublec: where's your blog?

00:45 <superherointj> bougyman, https://bluishcoder.co.nz/index.html

01:01 <bougyman> Thanks for a screenreader friendly blog, doublec :)

01:01 <bougyman> hating that pony moved to zulip. that interface is impossible to use for me.

01:02 <doublec> superherointj: I like Pony, it's a great language. It hasn't really come short for me. I would have liked to have seen the light dependent type stuff merged but I think it has bitrot too far.

01:02 <doublec> bougyman: I'm glad it works with the screen reader! Yes, it's frustrating so many things are using technologies that are anti-reader.

01:03 <superherointj> bougyman, I share the same opinion on Zulip. I think Pony could have both means of communication. The topic wording on this channel is quite harsh.

01:05 <bougyman> Yeah, i went and gave zulip a try. It took me 40 minutes and my wife's help to send a message.

01:05 <superherointj> doublec, you think Pony has bitrot? Or was it a typo?

01:05 <bougyman> that's just not tenable

01:05 <doublec> superherointj: that work that someone did for adding compile time light dependent types has bitrot

01:06 <superherointj> I see.

01:07 <doublec> superherointj: see https://vimeo.com/175746403

01:08 <superherointj> doublec, have you ever thought about improving it or contributing in some way that would make Pony have something nicer?

01:08 <doublec> superherointj: even so, I still like Pony. It doesn't reflect in my recent blog postings because I tend to get way too distracted by shiny features of languages and write about them haha.

01:09 <doublec> superherointj: unfortunately real world work and life is too busy for me to to contribute in that way lately

01:10 <doublec> superherointj: it takes a lot of time and patience to drive something through to completion such that it gets merged in most large projects.

01:13 <superherointj> Pony community is still quite small. Contributions are more easily accepted. Even if you don't take contributing too far, would be nice if you wrote about your views on it because you have a wide knowledge on languages and by making your objections public other people could improve on it.

01:15 <superherointj> Pony seems to be rewritten as Mare. Have you had a look on it?

01:16 <superherointj> https://github.com/jemc/mare

01:16 <bougyman> superherointj: hehe. I consider any project that's bigger than 1 pizza to be "big"

01:16 <bougyman> And I prefer small ones.

01:18 <superherointj> I have to remind myself of this quite often.

01:18 <superherointj> How do you define project size as pizza?

01:23 <bougyman> superherointj: how many people does one pizza feed?

01:23 <bougyman> Any more than that, and it's big :)

01:23 <bougyman> 2 pizza teams are still managable.

01:23 <bougyman> any more than that will lead to an eventual clusterfuck.

01:26 <doublec> superherointj: Mare is a slightly different language to pony I think, but using the pony runtime. Written in Crystal.

01:27 <superherointj> Different syntax but close semantics. Crystal as replacement for C.

01:28 <superherointj> doublec, compiling the objections you have on Pony might be relevant. Other developers could acknowledge that and even work on fixing it.

01:28 <superherointj> Even if you don't want to invest much time on it. At least Pony could be less of a bitrot.

01:29 <doublec> superherointj: I don't have any objections to Pony

01:29 <superherointj> doublec, if you ever find a language more interesting than Pony, that values correctness and performance and is fun (powerful) to write code on it, let me know. :-)

01:29 <doublec> superherointj: I'm happy with the way development is being managed and done. Except maybe not using irc :)

01:30 <doublec> superherointj: ATS is an interesting language if you feel like exploring correctness

01:30 <superherointj> doublec, I understood you were bothered about the way Pony dependent type support was being implemented.

01:30 <doublec> superherointj: I was sad it was not implemented. But it's still on their roadmap.

01:31 <doublec> superherointj: they don't need another "please prioritise this" which doesn't involve "and I will do it"

01:33 <superherointj> How well do ATL do in performance, concurrency? I found reference capabilities quite a nice concept to work with.

01:34 <superherointj> Actors + Reference Capabilities was a nice solution to the concurrency problem.

01:35 <doublec> superherointj: ATS doesn't really have a runtime and compiles to C so performance is on par with C. There's no concurrency libraries, you can use C threads or implement light weight threading. Its point of interest is the type systeem which allows low level type safety - including pointer arithmetic, dereferencing, etc.

01:36 <doublec> superherointj: yes, I like pony's actors and reference capabilities.

01:36 <superherointj> Should Pony be rewritten on it then?

01:39 <doublec> superherointj: it'd be an interesting project to do that but not something I'd recommend. ATS is too obscure and 'researchy'.

01:51 <superherointj> doublec, is ATS the most correct language you know?

01:55 <doublec> superherointj: I don't think there is a most correct language. Maybe Coq or Agda?

02:17 <bougyman> doublec: do you have a preferred web framework for pony?

02:17 <bougyman> I've been away from it so long I'm just going through all the tutorial/docs again.

02:17 <bougyman> and looking for ways I can use it at $dayjob

02:17 <doublec> bougyman: I don't think there is one

02:17 <bougyman> I found jennet

02:17 <bougyman> https://github.com/Theodus/jennet

02:18 <doublec> oh neat

02:22 <bougyman> I learn more about languages by reading code than I do reading docs, and web serving is easy to grok

02:24 <bougyman> this project seems to not follow what I read in the docs, though. At least, it has a Makefile. I thought one of pony's strengths was "no external build system necessary"

02:25 <bougyman> I really hate Makefiles.

02:25 <superherointj> Makefiles is simple.

02:34 <bougyman> doesn't look like the http server itself gets any love :( https://github.com/ponylang/http/issues/9

02:34 <bougyman> well, http, server and client

02:41 <superherointj> bougyman, when I don't find libraries in Pony I wrap something from C and it works fine.

02:42 <superherointj> It will take a while until mature implementations of most library needs appear.

02:43 <bougyman> as soon as you wrap C you lose pony's guarantees.

02:43 <bougyman> Don't you?

02:44 <superherointj> bougyman, the problem is restricted to the C library.

02:44 <superherointj> If the C library is good, your application will be good. And still worth doing this wrap.

02:45 <superherointj> It is of course better if everything was written in Pony. But from a practical POV I found this solution better.

02:45 <superherointj> Most times I find bugs in my application and not libraries.

02:45 <superherointj> And for that Pony is good.

02:45 <superherointj> At least it guarantees your application is fine, just not the imported libraries.

02:46 <superherointj> I think it is possible to import libraries from Rust too, but I have not tested yet.

02:47 <superherointj> If you want to be productive now, this solution of wrapping libraries is good. And when Pony ecosystem matures, you can always migrate.

02:47 <superherointj> Right now I am using SDL library wrapped.