During Saturday I attended the Scala LiftOff conference. THere were about 50-60 people there this year – many interesting people. The format of the conference was to have everyone propose sessions to talk about, and then put them in different time slots. This worked out great in practice, except for the small detail that the place we were in had terrible acoustics. It was extremely hard to make out what people were saying at times.
The exceptions to the unconference format were Martin’s keynote talk, before we started, and also something they called speedgeeking. I’ll talk more about that later.
So, Martin Odersky talked about the next five years of Scala – this information is pretty well covered in Dean Wamplers blog entry about the BASE meeting. I am impressed by some of the things they’re planning for the future.
After that I decided to attend a session John Rose put together, about JSR 292, invoke dynamic, and other features we could add to the JVM to make life for Scala easier. This turned into a pretty interesting discussion about different things. Martin Odersky was there and gave his perspective on what kind of features would be most useful. He was specially interested in interface injection and tail-call optimization, but we managed to cover quite a lot of ground in this discussion.
During the next slot I ended up being a butterfly – no session was really extremely interesting.
We had lunch and during that I saw Alex Payne describe some of the things they are doing at Twitter using Scala. After that came the speedgeeking. The basic idea was that twelve people should do small demos, max five minutes. They would do those demos for a smaller group of people, and then switch group – until everyone had seen those demos. I didn’t like this concept at all, and the way it worked out was just annoying – I ended up talking to John Rose and Martin Odersky for most of the time.
After that, me, Josh and Amanda figured that the weather was very nice so we moved our sessions outside. This also solved the problem with the bad acoustics. The first one of the outside sessions was Martin convening people to talk about equality and hash code semantics. The way implicits work right now make for some very strange and unexpected cases – such that “bob”.reverse == “bob” is not true. There are also several intricacies in how to handle hash code calculation for mutable collections. We didn’t really come to any conclusions, but Martin was happy that he’d gone through the available options and thoughts pretty thouroughly.
After that Josh and Amanda led a discussion about what kind of patterns we’re starting to see in functional object-oriented hybrid languages. Amandas experience with F# came in handy when comparing to the approaches used in Scala. No real conclusions here either, but lots of interesting discussions. My one mental note was to look up a recent OBJ paper, detailing an Object calculus. This reference came from Paul Snively.
After that the conference was over – me, Josh and Amanda was joined by Paul Snively for a beer. That ended up with me ranting about programming languages, as usual…
All in all, Scala LiftOff was a great conference, with a collection of many interesting people from several language communities. This ended up sparking very interesting discussions.
One Comment, Comment or Ping
Hi Ola! It was so nice to meet you and Josh and Amanda. What great fun!
Just to follow up on the OBJ comment: see . OBJ is actually a family of systems that sit at an interesting intersection of specification, implementation, and verification. I suspect, but haven’t really looked all that hard, that one could verify an object calculus such as with them. Or perhaps , which I’ve been trying to learn in fits and starts, would be a better choice. But the key observation to me is that for the longest time, there had been no formalization of “objects” at all. Now there are several!
June 9th, 2009
Reply to “Scala LiftOff”