o(ther)l/log

An interesting post reporting about the Lisp Conference 2005. Apparently there's some traction for evolutions and heavy experiments in the LISP world. Which is good.
It seems there are several efforts to implement Lisp-like languages on top of the CLR. Common Larceny is one of them: a R5RS Scheme implementation, with interoperability with other CLS languages.

What's notable is, just like IronPython, the performance is more or less is the same ballpark than 'classic' implementations. Worst on some micro-benchmarks, better on other. Of course benchmarks suck, and micro-benchmarks suck more, but this is a very good sign - at least to avoid the initial reject reaction.

I have to dig and see what other implementations of lisp-like languages exist for the CLR (I'm pretty sure there's one or two CommonLisp and I saw another Scheme implementation... maybe Dylan ?). It would be interesting to see some ideas of the functional/objects or actor languages experimented in the 80's again on modern infrastructure. While reading the slides, I was thinking about Ptitloo, the "object-scheme" hybrid language that Jacques invented after Mering. It could be a good M.Sc project :)