Monday, October 5, 2009

Some interesting thoughts from tim bray on 'the next java'. Having briefly played with Erlang, Clojure and Scala, it's interesting to see somebody else's take on them. Erlang was frankly just a little too out there for my tastes, although I can see the attraction (and I particularly like the assumption that everything is going to fail sometime). It just seemed like to much hard work to do useful stuff - for the things it was designed for it did seem to be excellent, though - particularly the definition of low level libraries. Maybe things have improved - I looked at it (and Mozart/Oz, briefly) about 5 years ago, and both seemed just a little too out there and difficult to integrate with anything not of their world.

Scala seems excellent, and has most of the things I look for in a language. I particularly enjoyed the Lift framework, and scalacheck. The big issue I've had is dealing with the tool support, which seems missing - although looking out there briefly now it seems like this might have improved, particularly with regards to building things. The actor framework seems nice in principle, but I haven't looked at it in much detail. It does seem relatively low level. There are some interesting changes on the horizon (continuation support as a compiler plug-in). I seem to keep getting half way through projects in Scala and abandoning them, though...

Clojure's a language I want to love, but somehow (like scheme, and common lisp), just can't manage to get my head around. Not sure if it's the size of my head, or if I just won't fit in the dimensions I have at my disposal, and I need a higher-dimensional head.... I do love the dynamic nature of it, but somehow I seem to be able to work better in rhino (javascript on the JVM).

No comments: