reactivation day!

Sigh

I'll leave you with a photo. Zach's right - turn off the bloody date display on your digital cameras!

There's not much more to say about ILC07. There are tons of comments floating around, and all agree on the value of getting together. I would recommend a reading of some other blogs to see how others viewed individual papers and issues.

Well, that was interesting...

So, we've come to the end of ILC07. I have come away with mixed emotions.

The organization and environment were wonderful. Nick L. deserves our gratitude for shepherding all his sheep and arranging the wonderful weather for this wonderful place.

The speakers were of variable quality, and of variable interest. Christophe Rhodes, in his talk "Extensible Sequences in Common Lisp" finished by exhorting the audience to bring Common Lisp into, at least, the 1990's. I don't think his tongue was completely in his cheek. He was wise enough to acknowledge the influence of a number of languages... even Python. Manuel Serrano gave a wonderful and impressive talk on HOP and the delivery of Web applications.

I thoroughly enjoyed my time at the conference and I learnt much, however I can't help but agree with Edi Weitz:

"(some incidents during the conference) ... confirmed my suspicion that there's a fraction within the Lisp community that is completely detached from what is going on nowadays. They did very cool things in the past, but then they somehow lost interest or they think that everything that came afterwards can't compare anyway so they don't need to bother."

A small group of us finished up at a lovely Thai restaurant, discussing Python, Erlang, Scheme and Haskell. Good ideas are coming from all directions. I hope to be at ILC08 and I'll pay my subs, but I'll expect the ILU board to have their organizational act together (not referring to Nick). It's 2007, and if you rest on your laurels too much, they get squashed.

Dinner last night

The Sidgwick Site Law Faculty building; full of Lispers!

Dinner last night was great. Interesting company (a "quant" who actually was a particle physicist, Alexander Repenning and the MD of Genworks were amongst the crowd). Lots of talk about Common Lisp, the evils of smoking, the coming of Erlang, the tragedy of MCL, and maybe new things from Allegro. Got back to Clare around midnight.

So, onwards and upwards!

What were they thinking?

End of Day 1 presentations, and on the way back to Clare what do I see? This most foreboding building. In a place of such beauty what where they thinking? This is the Library of Cambridge University; one of the great libraries of the world. My deep aesthetic revulsion was confirmed when I was told that the building was used as a location for the Thought Police in a film version of 1984.

A group of us are going our to dinner. I'll try and bring some back in a doggy bag and if Zach scratches his screen hard he might get a taste...

Some compensating beauty:

Just for Zach

The Mathematical Bridge



The Bridge of Sighs



Wren's Library at Trinity

The flickr tag for the ILC07 photos is ilc07 (according the he-who-must-be-obeyed N. Levine), but my many :-P pics wont be available until flickr.com check my (recently opened) account to make sure that I'm not posting naughty things.

Today's sessions have been a blast, with Peter Herth telling us (and showing us) Ltk; a real solution to a real problem - the "lightweight" GUI toolkit for CL is now a reality. It's now possible to do a GUI based Hello World in Lisp at a level that all newcomers can understand and use.

Cyrus Harmon gave a terrific talk about the use of CL for the analysis of spatial patterns of gene expression in the fruit fly. It did my heart good (being an ICU doc) to see some hard core biology + hard core lisp.

Some more thoughts on the tutorial day

(not cambridge ... :)

Well, having had some time to think over the all day Common Lisp tutorial, I've come to some to some tentative conclusions. Most of the people in the room were experienced programmers (physicist, software engineer, etc), and were very clear about the current "conventional" programming models. When the use of with-open-file was discussed, the instructor was surprised that such finalisation facilities existed in other languages. Of course, the point is that this facility was introduced into CL without needing to wait for the "next version of the compiler", via (ta-da) macros. Rather than be told the search-time advantages of hash-tables over linear lists, it would have been more interesting to see why the at-first-glance clumsy hash-table access (setf place value) was more broadly useful, and move into a discussion of "places" and then defining your own setf-able stuff.

A lot of time was spent on a text processing example which would have been much more succinctly expressed in Perl (or Python, or Ruby, or whatever). With Smalltalk, Ruby and Python all having (different) meta-programming facilities, it would have been good to see how MOP worked. Finding the class, slot-names and number of slots of an object at run-time is not unique to CL. Time ran short in the end, so there was little time to talk about macros.

The program made it clear that "if you already know Lisp then this isn't the place for you." Most of us could were looking for examples of "Why Lisp moves you past whatever you're already using". I might put some suggestions together for next time...