April 25, 2009

European Lisp Symposium 2009 - programme details!

On May 27-29, there will be the 2nd European Lisp Symposium taking place in Milan, Italy. The program looks very exciting, and I will definitely be there (well, due to obvious reasons, see below ;).

For example, there are two keynote talks. One is by Kent Pitman who is an award-winning author of technical papers about Lisp, editor of the ANSI Common Lisp specification, and designer of the HyperSpec, the de-facto standard manual for Common Lisp. He will discuss how the Lisp community should move forward from his perspective. Kent's ideas are always well thought-out, although at times controversial, so this is certainly going to be a provocative talk.

The other is by João Pavão Martins and Ernesto Morgado, the two owners of SISCOG - a Portuguese company that develops large-scale industrial planning and scheduling software which is in use for over twenty years. It's always good to hear what practitioners have to tell about their experiences with Lisp, so this should turn out quite interesting.

There will be a couple of presentations in the main track of the symposium about papers that have been reviewed by a program committee chaired by António Leitão.

Jim Newton is going to present a type inferencing approach for the Skill dialect of Lisp that is actually being used in his group at Cadence Design Systems, one of the world-wide largest providers of Eletronic Design Automation. (If you use an eletronic device, it's very likely that the chips inside were designed using one of their tools!)

Thomas Burdick is going to present an approach for compiling FEXPRs (think: first-class macros that you can pass around like regular functions). I'm a bit skeptical here that this will work, because it is actually known that FEXPRs cannot be compiled, but maybe he has found an interesting twist to the problem.

Some colleagues of mine from the Artificial Intelligence Lab of the Vrije Universiteit Brussel in Belgium are going to present a debugging technique they have devolped for their own large-scale agent system that they use to investigate the possible development of natural languages by reconstructing how they could have evolved from simple, first-class principles. The debugging approach consists of monitoring the activities of the agents and presenting the process as an interactive webpage that can present arbitrary detailed (or abstract) views on what is happening. Very cool stuff here, and since they use web technology here, it is actually portable...

Charlotte Herzeel is going to present an architecture for Software Transactional Memory (STM). STM has become quite a hot topic in the last few years because it seems a very promising approach for dealing with concurrency, and since multicore processors are the buzz of the moment, there is a lot of interesting in such approaches. However, little attention has been payed to the design of STM frameworks where you can selectively plug in different STM algorithms. Charlotte has developed a reflective approach (think: metaobject protocol) for STM, which seems very promising (but I'm one of the co-authors of the paper, so I'm naturally biased, of course ;).

Another presentation will be about Linda-style distribution layer on top of Kenzo, an apparently very powerful system for symbolic computation developed in Common Lisp. The Linda Model (also used in JavaSpaces and TSpaces, for example) has a number of very interesting properties for distributed computing, and the paper presents an implementation based on AllegroCache, a robust and high-performance object-oriented database for Allegro Common Lisp. I'm wondering what the concrete benefits for a symbolic algebra system are, so this is another presentation to look forward to.

Finally, I am going to present a paper myself - that's the main reason why I will definitely be there ;). My presentation will be about a macro system on top of Common Lisp's macros that allows writing hygienic macros. The system provides facilities similar to Clinger's renaming construct. The interesting part is that my system is implemented in fully portable Common Lisp and integrated in such a way that 'regular' Common Lisp macros and the macros developed with this new macro system can be used together. The reason why this works is because of the use of symbol macros in the implementation of the new constructs.

Other items on the programme for the symposium are:

A debate on expanding and updating the Common Lisp standard. I am known to be quite conservative when it comes to this topic, in that I don't think ANSI Common Lisp needs a serious revision, but can be developed in a piecemeal fashion (and this actually already happens due to the efforts of the currently very vibrant Common Lisp community). I believe (obviously) that something like CDR is much more promising than a "big bang" revision of the core language. The recent developments in the Scheme community, where the highly controversial R6RS specification was almost not ratified, seems to indicate that the danger is too high that a lot of time and resources could be wasted that can otherwise be used in a much more productive way. Well, maybe there will be some new ideas and visions coming out of the debate...

On the Saturday after the symposium, there will be a visit to the Futurism exhibit in Milan. Futurism was an art movement in Italy in the early 20th century, whose founders announced that everything "old" (so artistic and political tradition) should be replaced by the new ideas of a then young generation. This is probably one of the strangest choices for the social programme of a Lisp-related event: Lisp is second oldest programming language still in use today, and has always been in competition with whatever other programming language came along that was considered to be 'newer' (as if that automically meant 'better'). However, judging for example from this blog posting, this could actually be a very inspiring exhibition.

So all in all, a highly interesting programme, with some more items being added in the coming days. (There are rumours that Christophe Rhodes is going to give a tutorial about non-portable features of SBCL, for example!) Although the early registration deadline is ending very soon now, there is no reason to despair: With €80 for students and €160 for regular participants, the registration fees will remain very low.

So, hope to see you in Milan!