Keeping Lisp alive and practical
ilc07 If you need someone to build you an extensible, elegant and demonstrably functional application in a hurry then the Law Faculty at Cambridge University is currently the place to be.
The building, one of the modern steel and glass blocks that nestles close to the much-reviled Seeley History Library on the university’s Sidgwick site, is currently home to the 2007 International Lisp Conference, and the lower ground floor is thronged with developers who like to create the ‘ill-defined, complex and dynamic’ applications that Wikipedia claims Lisp is good for.
Most of the hundred or so delegates are men, many of them would seem to be of an age to remember the good old days of minicomputers – if not mainframes – and enough have long hair and beards to make your correspondent feel at home, but this shouldn’t detract from the importance of the work that takes place in Lisp or the enhancements they are making to this much-loved language.
Invented in 1958, Lisp sits with COBOL and FORTRAN as one of the oldest languages still in regular use for production systems, and over the last decade Common Lisp has become a hidden jewel for developers who need to build flexible applications in a wide range of domains.
Those who arrived in Cambridge on Saturday got a beautifully sunny afternoon for their introduction to the city, with a punt trip on the Cam to get them into the mood for four days of tutorials, presentations, networks and parenthesis counting.
Keeping the language alive means investing in training, and Sunday was dedicated to practical work, with a range of tutorials for those who wanted to learn more about the language, but the real action started today with a brutally early 0830 kickoff.
Despite the hour the room was full for Herbert Stoyan’s romp through the history of the language, demonstrating the sort of dedication that has made Lisp a real survivor.
Talks range from the discursive to the entirely practical, like Peter Herth’s presentation of how to use LTk, a Common Lisp binding for the Tk graphics toolkit, which included code fragments and onscreen editing, while Robert Strandh showed us how to re-implement EMACS in Common Lisp, primarily just to show that it could be done.
The variety of talks is partly because this is neither an academic conference nor an industry event designed to sell tools but a cross-community event.
Christophe Rhodes, from Goldsmiths College in London, was pleased to see such a wide range of people, ‘from hobbyist to industrial users to academics’, as he put it. ‘They come together in a way that you don’t get at university conferences, where you just have academics,’ he said. ‘Here we have users who use Lisp as a tool, as well as people who do language development.’
The view was reinforced by Marie Burton-Aimar from the EPITA research and development lab, who added that ‘there are two generations of Lispers – the old ones who been programming for twenty years and those new to Lisp’. She felt that bringing them together in one conference was good for the long-term prospects of the language, even if there were differences of opinion and approach.
As well as bringing the community together, the conference includes the announcement of the results of the Lisp programming contest (covered in El Reg on Feb 3). Entrants had to write a program to play the game of ‘Continuo’, a card game where players place coloured cards on the table and score points. Each card has a 4 by 4 grid of coloured squares, and you score points by building connected regions.
The first entry arrived the day after the competition was announced, and prizes were distributed as liberally as parentheses in a badly-written Lisp program. Christophe Rhodes got one for obscure use of Lisp for ‘relying on the non-nil return value of (digit-char-p)’, while Frank Buss got the ‘first half-decent entry’ award and Guido Witmond won for learning the language just in order to enter the competition.
Conference organiser Nick Levine,who lives locally and works at software house Ravenbrook, sees the event as a way to bring the two elements of his life together. "I’ve made a living from Lisp for almost twenty years and lived in or around Cambridge for nearly thirty", he says. "Now I can bring the two together – even if organising the conference has taken over my life!"
The conference continues until Wednesday, finishing with Jans Aasman’s paper on scalable Lisp applications, while delegates can also look forward to talks from community superstars John Mallery from the AI Lab at MIT and Ralf Moeller from Hamburg University of Technology.
Another highlight will be Manuel Serrano’s attempt to bridge the cultural divide between the sort of applications we typically associate with list processing and the brave new world beloved of the venture capitalists with HOP, a Lisp-based environment for developing Web 2.0 Applications. It could get messy in there, although at least the parentheses will balance. Further details on the ILC 07 blog. ®
Sponsored: Hyper-scale data management