❄️🦊 @icefox@octodon.social

ahahahaha gitlab.com is getting slammed:

https://status.gitlab.com/

@jelle_dc It does seem the best bet for it, yes.

@charlescodes Git is decentralized but has no notification system and API to let programs using git to communicate with each other in a consistent fashion. Git's own decentralization model is built around emailing patches around. Which is fine, but very primitive compared to Github.

Gitlab doesn't federate, afaik. If I run my own gitlab server and want to make a PR to your Gitlab server, I need to make an account on your server.

@jon_valdes The traditional payload is a pointer. I don't know if LuaJIT invented the technique but I'm pretty sure it popularized it.

So when are we going to collectively suck it up and write a federated GitHub-like thing? I mean, we've been needing it for years and just nobody's gotten around to it.

@dasyatidprime Me too, somewhat. That's one of the nice things about collaborators, it is easier for me to be excited by a project when other people are working on it too.

Working on ggez is making me realize that making a release is at least partially a matter of deciding which bits to procrastinate on until the following release.

@susannah Fair enough!

@susannah Why do you need to CW food stuff? 😕

@pfm @phoe And I said "becoming obsolete". :-P This doesn't mean they're going away, it means they're starting to look a lot more like statically-typed language. Gradual typing is part of Julia and is now infecting Python for example, and probably Javascript soon (dunno about Ruby)

So my prediction, based off of current data, is in 10-20 years purely dynamically typed languages will be considered as old-fashioned and ill-principled as a large software project without tests is considered now.

@pfm @phoe

More expressive than earlier static type systems. Consider C or Pascal's type system, then consider C#/Java, then consider Rust's, then consider Haskell's. The amount of metaprogramming and reasoning about types and their properties gets increasingly more sophisticated, so the guarantees you can make also gets more and more comprehensive. Now you have sum types, product types, dynamic dispatch, generics, higher-kinded data types, and so on. C has `void *` and that's all you get.

@phoe Yeah, I'm aware of that. The Python people at least call it "gradual typing" and it's really nice, but CL naturally had it like 30 years before anyone else.

My main priority is the safety rather than the performance, to be honest

...by puking on the carpet it seems.

Oh well, c'est la vie.

My travels are done,
To and from the sun.
No matter where I roam,
My cats welcome me home.

Am I the only one who feels like dynamically typed #programming languages are becoming obsolete? Type systems are more expressive and type inference is more ergonomic, so there seems to be few arguments in favor of weaker typing.

@kara I reeeeeally wasn't thinking that far ahead! Lots of water passed between then and now. I shall ponder.

But there's often a rise-and-fall pattern of empire that is much more apparent in Asian and Middle Eastern empires than European, where it lasts 3-5 generations before falling apart. Moguls and other Gunpowder Empires, certain caliphates, others recognized by Ibn Khaldun even in the 14th century. The Big Important Empires are the ones who avoided the typical generational sine-wave.

@kara There seems a very strong self image of what "a Roman citizen" should be, and a surprising amount of formalism and bureaucracy around that idea from an early time, especially for 400 BC Italian city states. Those traits were quite militaristic, engendering conquest, and successful conquests this self-reinforced this ideal. They ended up with a weird balance of liberty and convention and organization that many empires did not; id est, the Moguls.

I am not an expert on the topic, I confess.