While day 1 focused on invokedynamic day 2 focused on time and locality.
Very thought provoking. Argued for immutable data structures and pure functions, but with time transitions and observers. Nice pragmatic response in a way to Erik Meijer's Fundamentalist Functional Programming from last year's JVM language Summit. This was somewhat a philosophical presentation, but much like last year it made me want to re-visit Clojure. Here is a better summary than I could have written: http://wiki.jvmlangsummit.com/Clojure_Keynote
Cliff Click
Awesome scheduling to put this immediately after Rich's talk. Cliff offered up four presentations and the audience overwhelmingly chose the one about x86. Interestingly this had some similar theories to Rich's talk (time and locality), but at an extremely low level.
Hotswap
Research presentation from European students that have partnered with Sun. This was interesting, but the limitations make it not that useful for real production code hotswapping in my opinion.
Groovy
This talk by Theodorou focused on the performance of Groovy and was almost an advertisement for not using Groovy. Of course the performance of Groovy doesn't match Java, but it is really bad in some instances. The moral was use Java where performance is necessary.
Sun update
This was probably not recorded, but Octavian took hard questions from Josh Bloch, Neal Gafter and others. Hopefully the departure of key Sun employees ends up helping Sun in the long run. In my opinion Sun is overestimating the value of the Java store.
Erik Meijer
Not nearly as entertaining as last year's talk, but still interesting and based on mathematics. Erik was looking for solutions to concurrency and resource cleanup problems that are difficult and that currently have no "clean" solution.
JRocket
Frederik's talk focused mostly on anti-optimizations and basically asked the language designers to not try to implement optimizations that are also performed by the JVM. The audience was hoping for specific patterns more than Frederik offered up, but Frederik is clearly very knowledgeable about JRocket and the JVM specification.
Ioke
Ola tried to give a one hour introduction to Ioke in 30 minutes and by his own admission it did not work out that well. Many in the audience were not familiar with the syntax of IO which originally inspired Ioke and that did not help him very much. The BNF was shown much more than actual source examples.
Dinner
Dinner was at Faultine which has good food and terrible acoustics. Aside from having to scream at the people sitting near me it was a brilliant time. Dick Wall of the Java Posse was very friendly and a fantastic ambassador for the Java/Scala community. If you aren't familiar with the Java Posse podcast then I suggest you check it out on iTunes or at http://javaposse.com/. Their google group and mailing list are also interesting.
0 TrackBacks
Listed below are links to blogs that reference this entry: JVM Language Summit 2009 - Day 2.
TrackBack URL for this entry: http://www.nearinfinity.com/mt/mt-tb.cgi/635



Leave a comment