PL folks: Steve claims there it's been 10 years since we've had interesting new languages. Please provide counterarguments. - ThreadSky

shriram.bsky.social • 1 day ago

PL folks: Steve claims there it's been 10 years since we've had interesting new languages. Please provide counterarguments.

Reposted from Steve Klabnik

Speaking of which, I’ve been wondering when the next new batch of interesting languages is gonna appear. Like it’s been ten years since the go/rust cohort came up. A lot of current up-and-coming languages have been up-and-coming for a while, when are we gonna see some fresh faces?

Comments

psteckler.bsky.social•1 day ago

Does OCaml 5, with multi-core support and effect handlers, count for anything? Those new features certainly make the language interesting.

shriram.bsky.social•18 hours ago

Yeah, great point. Just because the *language* isn't new doesn't mean there aren't a bunch of interesting new *ideas* in it. The constancy of the name badge hides a lot of things. (For instance, how Scala is incorporating capabilities.)
CC @steveklabnik.com

steveklabnik.com•18 hours ago

Yeah for sure, the latest OCaml stuff is very interesting!

chriskrycho.com•1 day ago

To be fair, I think Steve meant “Go and Rust are quite old, and the up-and-coming ones like Unison, Zig, etc. are *themselves* half a decade old, so what’s new in the last year or two?”

zee.town•1 day ago

Though I get Steve's intention in the og post, there should be more discourse on the topic. With Strange Loop "over", I think live chatter is lacking in the PL industry/research cross-section. I learned about Rust & Julia back in 2012 b/c of the emerging langs workshop: https://github.com/nyergler/strange-loop-2012/tree/master/emerging-languages.

shriram.bsky.social•1 day ago

This is a great point. There are various other conferences but nothing with the gravitational pull of StrangeLoop. Plus, I think the generous travel $ of the 2010s are going to be mostly gone. I definitely viewed StrangeLoop and the other conferences of that era as another ZIRP phenomenon.

bauer.codes•5 hours ago

The 2012 SL emerging langs workshop is one of my favorite conference events of all time. ❤️

zee.town•4 hours ago

For real. That was just so amazing, even more so in retrospect.

chriskrycho.com•1 day ago

Ah, great insight about StrangeLoop, which is also what launched Elm to whatever passes for prominence in PL nerd circles, and then indirectly a bunch of related/downstream tech. We really need a StrangeLoop vNext. 😮‍💨

zee.town•1 day ago

Yep. More than Elm, it also hosted many academics in PL and systems, esp. in earlier years. Hanging out with Prof. Dan Friedman early on had a major effect on me for example.

shriram.bsky.social•1 day ago

Hanging out with Prof. Dan Friedman early on had a major effect on me, too. (-: (And I wasn't even "hanging out", I was just a total stranger yeeting email at him and he was chill with that, what a mensch.)

zee.town•3 hours ago

Yeah. Really amaze times. The icfp+SL year included SPJ joining me at a table in the hotel lobby while I was trying to prove something related to church-rosser for Frank’s Tapl course. Grand moments.

bandukwala.me•1 day ago

That’s how I found out about Ravi Chugh and Cyrus. Rip strangeloop

shriram.bsky.social•1 day ago

Don't think I interpreted it differently. I'd say in return, and I mean this genuinely, "Which conference proceedings have you looked in?" That's obviously not at all the only place we can publish new languages (thank goodness), but it's certainly where you'll find many novel things (by definition).

ltratt.bsky.social•1 day ago

A challenging question for us researcher is: when does a language become interesting? I know a lot of people think "when I've published a paper on it (w/basic implementation)" but I'm no longer sure that's enough. Only at scale do many problems and opportunities become truly apparent (c.f. Rust).

mvsamuel.bsky.social•1 day ago

Yeah, the efficacy of things like Rust's error message diagnoses of borrow problems only become apparent after the user base has grown past new language early adopters to generalists.

ltratt.bsky.social•1 day ago

I evaluated Rust twice, but only on my third go (late Dec 2014) could I convince myself that the cute ideas were likely to be usable and hence were (to me) worthy of thinking of as genuinely interesting. A few days later my research group started our (long) conversion to using Rust by default.

ltratt.bsky.social•1 day ago

A few things stand out for that from me. It wasn't a new language by the time of that third evaluation, and had quite a bit of effort put into it. But, even so, it still required me guessing/hoping that its trajectory would make it the "next big language". I'm lucky my guess was right!

krismicinski.bsky.social•1 day ago

I have had a lot of fun developing our language ideas myself--but I would never want to compete with a general-purpose programming language. Can you imagine what a pain that would be, and how much expertise it would require? So I work on trying to make small contributions that infiltrate ~Rust

ltratt.bsky.social•1 day ago

I once thought I could. I learnt a shedload, technical and non-technical! There's definitely a wide spectrum of worthwhile possibilities!

krismicinski.bsky.social•1 day ago

PL isn’t just about building new languages. In the same way that CS isn’t about just computers. It’s a broad set of methodologies used to explore computation.

shriram.bsky.social•1 day ago

I mean, we've been working on Pyret for a decade now…

daveandersen.bsky.social•1 day ago

Yeah, but are they publishable, or do they intersect with interesting and useful?

shriram.bsky.social•1 day ago

Go troll your CSD buddies, Dave.

daveandersen.bsky.social•1 day ago

No way, they know where to find me in person

steveklabnik.com•1 day ago

i think i wasn't super clear in my original post, partially because i was just offhandedly saying something that was on my mind. oops, haha

steveklabnik.com•1 day ago

i did mean it in the sense of like, "breaks out into industry where you can start getting jobs." obviously PL journals are releasing tons of new and interesting stuff, I'm sure, but that's not the maturity level I was thinking of.

steveklabnik.com•1 day ago

totally separately from the thing i was musing about, i do like hearing about more niche things though. i am certainly not knowledgeable about the actual state of the art here; a past me may have wanted to go into academic PLT, but present me just doesn't have enough time to do everything :(

zee.town•1 day ago

just do what i do..., wait it's 4am and I'm up every night, nope, don't do it haha.

nsri.bsky.social•19 hours ago

Mojo Lang. From Modular. From Chris Lattner. Based on MLIR. Unified - I.e code can run on CPU, GPU, FPGA, AI ASICS, custom silicon etc.

gjmichaelson.bsky.social•1 day ago

There have been no new semantic abstractions for decades. Only tweaks to syntax, analyses, run times and libraries.

shriram.bsky.social•1 day ago

Rust patently presents a new semantic abstraction.

gjmichaelson.bsky.social•22 hours ago

That's core Rust as opposed to the standard library?

shriram.bsky.social•18 hours ago

Yes, ownership is not part of the "standard library", it's imposed *on* the standard library. You know this, so I'm baffled by your question.

gjmichaelson.bsky.social•17 hours ago

I phrased that poorly. The Rust standard library exposes the memory model. And unsafe undermines it. Would you argue that the Python memory model represents a semantic innovation?

hjvt.dev•17 hours ago

Does unsafePerformIO undermine Haskell's aspirations of being pure?

shriram.bsky.social•13 hours ago

I don't know where your Python question is going, but I wouldn't have said it's a semantic innovation. (Lua maybe.) But the unsafety doesn't preclude the innovation of Rust.

Comments

Posting Rules

Reply