I just got back from the unbelievably awesome Emerging Languages Camp at OSCON. I wish I could come up with a good way to get across how cool of an experience it was. All I can really say was that for the first time in a long, I felt like I was really around my people.
A programming language nerd is a really specific kind of nerd and I don’t really know many who share my obsession. For the last two days, I was in a room packed with people who did.
It’s hard to summarize two intense days of full-on nerdery, but I thought I would talk a bit about something good I got from some of the talks. This doesn’t mean I agreed with them all (honestly, I found the Go talk myopic), or that anyone I omitted here wasn’t interesting. I’m honestly exhausted this morning, so I only have so much steam.
Rob Pike (Go)
The theme that comes up frequently when I read about Go or hear Rob talk about it is this sense of pragmatic feature economy. Any time someone asks, “why doesn’t Go have Foo”, his answer tends to be “We found in practice that we don’t miss it.”
Language designers can tend to be magpies, wanting to cram in all of our favorite features just because they’re shiny or to have a longer bullet list than the other guy. If those features don’t help real people solve real problems, you’re just wasting your time and making things more complex for them. While I don’t always agree with the specific choices that the Go designers make, I do agree with the “let’s see if we can get by without them” approach that leads to them.
Phil Mercurio (Thyrd)
Phil’s talk was one of a few on visual programming. Most of my experience with visual languages is that they’re either toys or incomprehensible timecube- esque crankware.
Phil got up and humbly walked through his visual language. When he asked for questions at the end, there was a long pause before any hands went up. From talking to people afterward, I think the consensus was that we were all thinking, “Holy crap, he actually did it.”
He managed to find a small kernel for a visual language (nested grids and triads) and build something out of it that really works. It’s really hard to pull something like this off, and I think it’s cool that’s he had the insight and tenacity to keep cranking away at it.
Alan Eliasen (Frink)
Alan stressed that Frink is a general-purpose language with user-defined functions and stuff but its raison d’etre is doing arithmetic with units, which it does so unbelievably comprehensively that it boggles the mind. From his talk, I can only assume that Alan’s house is packed floor-to-ceiling with yellowing almanacs and dusty IEEE standards reports. Calling his knowledge of measure encyclopedic is giving encyclopedias too much credit.
What I found interesting about Frink is that by limiting its focus to a single domain, Alan has ended up with a language that’s more powerful and useful. When asked what their language is good for, many designers would say “everything” which really means “nothing”. Frink embodies the principle of doing one thing and doing it well.
Gilad Bracha (Newspeak)
The camp was an even mix of academics and hackers. A lot of academic talk is over my head so the mortarboard team is underrepresented in this post. Gilad’s talk was in that category: a lot of it was too high-level for my primitive monkey brain to grasp.
One thing really stuck with me though. Newspeak doesn’t have namespaces, modules, or a lot of other OOP features. What it does have is classes, and classes nest. That simple core is flexible enough to do duty as a module system, an abstraction system, a namespace system etc. I love the idea of getting one simple concept and letting it fully unfold to cover a bunch of different problem solutions.
Jeremy Ashkenas (Coffeescript)
I totally didn’t pay attention to Jeremy’s talk because I was busy slapping together slides for mine. This sucks because I think Coffeescript is fantastic. I love its obsession with aesthetics and clarity. Both the language and its documentation are a testament to the philosophy that beautiful is usable. I wish more language designers would pay heed to that.
Steve Folta (Trylon)
Steve’s talk embodied two ideals that I think are often overlooked. The first thing he said when he got up was, “Trylon doesn’t do anything new.” That puts Trylon in the same space as Python, Ruby, and almost every popular language today. If you want to publish papers and advance the art, breaking new ground is key. If you want a language for people to use, there’s real value in staying away from the borderlands.
The other thing he said that struck me was, “When I get home I program in Trylon because it’s the language that feels the most fun to me.” A well-worn truism in writing is “write for yourself.” Trying to make a language that’s all things to all people is futile and exhausting. I love that Steve just made the language that he wanted.
Amos Wenger (ooc)
Any domain has two levels of knowledge: the core ideas for the domain, and the cultural wisdom around those ideas. The first tells you how to do stuff. The latter often tells you which stuff not to do. Any well-versed programming language person can tell you about both recursive-descent parsers and generated parsers. They’ll also tell that generated parsers are the “right” way to do that.
Most of the time that advice will save you from wasted effort, but sometimes I think it keeps people from going down paths that may actually be fruitful. Sometimes the thing that everyone knows is true isn’t. (For example, every language I know of with a lot of real-world users actually does use a hand- written parser.)
Amos is a young French iconoclast. If he’d been born in a different time, I expect he’d be a brick-tossing anarchist. One advantage that attitude gives him is that he and the others working on ooc pour features into the language while the rest of us are still sitting around fretting about minutia. I think a lot of us could use some of that “let’s just fucking do it” spirit.