Photos are available at http://dept-info.labri.fr/~strandh/LSM2004/, courtesy of Aleksandar Bakic.
The LSM is an inexpensive and informal event. The organization tends to be a little haphazard, but this is compensated for by a congenial atmosphere. Cheap accommodation in almost salubrious University dorms is available, or there are plenty of hotels downtown. Bordeaux is 3 hours by train from Paris, and good food and wine abounds.
See http://wiki.ael.be/rmll2004/index.php/ThemeLangagesDeHautNiveauPourApplications and https://abul.org/Rencontres-mondiales-du-Logiciel,135.html.
Things to do
There has been some discussion on irc about whether we should get together and hack collectively on some project, as opposed to 2002 where people tended to work on their own individual projects. Two possibilities have been advanced so far
- Frustrated by the recent poor security record of CVS (and, really, anything else written in C that listens on a network port), we thought about implementing the Arch version control system in CL, starting from the specification (I hope there is a specification ...) (There's a specification for the changeset format; implementing commits and so forth off of that doesn't seem terribly difficult. The tricky part, of course, is starting to work on merging ...).
darcs also provides an interesting alternative to cvs and arch, with a well thought out patch model suitable for distributed development and since its written in haskell it cant be all evil.however, darcs doesn't have a specification, which is what is interesting here. - The alternative, which admits a greater chance of cool screenshots, is to create the currently mythical CL-Emacs. Daniel Barlow's suggestion was to start with portable hemlock, work out how to install it in a selection of free CL implementations, add features to taste, and graft the SLIME protocol onto it (replacing WIRE, which is how unportable Hemlock currently talks to CMUCL)
(In fact neither of these happened to any great extent, but we did admire Gilbert's CLIMified Portable Hemlock)
Lightning presentations
The format of a lightning presentation is probably negotiable, but in essence is as follows: 5 minutes. There will be time afterwards for questions/discussion; how much time depends on how we go. Informal proceedings are available.
Sign up here if you can give one or two:
- Daniel Barlow: using asdf and asdf-install for your projects
- Daniel Barlow: SBCL's threading interface
- Rudi Schlatte: simple-streams: good, bad and ugly.
- Rudi Schlatte: xml-rpc.
- Christophe Rhodes: Common Lisp and :IEEE-FLOATING-POINT.
- Tim Moore: CLIM and multiple applications. There was some discussion on this, and whether/how the command loop should be shared, back in the early 1990s, roughly between CLIM 1 and 2. See the 1992 section (GZip archive) of the CLIM mailing list archive, and possibly later sections (see the threads starting with messages 9207131510.AA01715@jonquil.uchicago.edu -- Subject: Overlapping windows to CLIM application frameworks; and 19920715132226.2.KEUNEN@nrbmi1.ia.nrb.be -- Subject: proposals for meta-applications, frameworks, windows...(long message)) -- Paolo Amoroso
- Andreas Fuchs: Making CLIM applications
not use courierpretty. - Gilbert Baumann will try to sell Portable Hemlock.
- Bruno Haible: using the MOP to implement an FFI to Java.
- Bruno Haible: Case-sensitive Common Lisp packages.
- Eric Marsden: Effective use of SERVE-EVENT.
Attendees
- Daniel Barlow
- Christophe Rhodes
- Andreas Fuchs
- Rudi Schlatte
- Gilbert Baumann
- Deliana Foutekova
- Eric Marsden
- Bruno Haible
- Tim Moore
- Aleksandar Bakic
- Robert Strandh
- Arnaud Rouanet
- Matthieu Villeneuve
- Nicolas Lamirault
- (add your name if you were there)
Topic: Conference