exit/1

April 20, 2019 9:29 AM

“Make it work, then make it beautiful, then if you really, really have to, make it fast. 90% of the time, if you make it beautiful, it will already be fast. So really, just make it beautiful!” Joe Armstrong, inventor of the influential (and beautiful) Erlang programming language, has died.

Erlang: The Movie

The Mess We’re In
Learn You Some Erlang For Great Good

posted by swift (26 comments total)

.
posted by a snickering nuthatch at 9:50 AM on April 20, 2019

I had to learn some Erlang the hard way ages ago. I wanted a chat server for a language learning site and chose XMPP and ejabberd but wanted to hack it a bit to add a bunch of features. Funny thing is that once I started getting error messages and stack traces I had an epiphany... those were the previously inscrutable error messages that sometimes (rarely) spat out of my work's network monitoring application. Anyway, after the first chat server I went and set up a networking/security chat server for work that's been running for fifteen years or so without a problem on a machine that stays up for years and years at a time.

Erlang is pretty cool that way, stable as a rock.

.
posted by zengargoyle at 10:16 AM on April 20, 2019 [1 favorite]

.
posted by JohnFredra at 10:19 AM on April 20, 2019

.
posted by Alterscape at 10:24 AM on April 20, 2019

One of the few languages I've seen that actually uses '.' to terminate a declaration.
posted by atoxyl at 10:37 AM on April 20, 2019 [5 favorites]

.
posted by adamsc at 10:41 AM on April 20, 2019

.
posted by techSupp0rt at 10:45 AM on April 20, 2019

.
posted by sammyo at 10:55 AM on April 20, 2019

.

I found Joe in Twitter about two years ago and I was taken with how he affably interacted with developers and admirers around the world.
posted by putzface_dickman at 11:14 AM on April 20, 2019 [1 favorite]

.
posted by northtwilight at 11:31 AM on April 20, 2019

.
posted by acb at 11:44 AM on April 20, 2019

.
posted by grimjeer at 12:08 PM on April 20, 2019

.
posted by Anonymous Function at 1:21 PM on April 20, 2019

.
posted by weft at 1:24 PM on April 20, 2019

.
posted by bouvin at 2:04 PM on April 20, 2019

Hello Joe. Hello Mike.
posted by Nelson at 2:31 PM on April 20, 2019 [1 favorite]

.
posted by The Power Nap at 4:28 PM on April 20, 2019

.
posted by Quackles at 5:00 PM on April 20, 2019

.
posted by mantecol at 6:40 PM on April 20, 2019

.
posted by I'm always feeling, Blue at 7:34 PM on April 20, 2019

.
posted by Cash4Lead at 7:44 PM on April 20, 2019

.
posted by suetanvil at 10:16 PM on April 20, 2019

.
posted by osi at 8:03 AM on April 21, 2019

I’m tempted to propose a rule similar to Greenspun’s tenth that posits that any sufficiently complicated multithreaded program contains an ad-hoc, informally-specified, bug-ridden, slow implementation of half of Erlang, because I’m quite sure that I’m not alone in having come to the conclusion early on in writing multithreaded programs that I could avoid a lot of pain and suffering by avoiding shared memory and instead forcing threads to communicate by putting messages on each other’s thread-safe queues and only taking action on their owned resources in response to those messages.

Of course, that’s only a small part of what Erlang’s execution model brings to the table, and of course there are performance-sensitive applications where that model doesn’t suffice, but it was a nice moment when I learned that I wasn’t barking up the wrong tree entirely.

. for a great thinker and educator.
posted by invitapriore at 3:01 PM on April 21, 2019 [5 favorites]

Also, I know it’s Hacker News, but this thread has an account by someone who emailed him out of the blue when they were still green about concurrency, and further down are Armstrong’s actual responses, which are wonderfully gracious.
posted by invitapriore at 3:33 PM on April 21, 2019 [3 favorites]

.
posted by silentbicycle at 5:20 PM on April 22, 2019