Still struggling with physical side effects of the conference. It's something of a cliche to say so, but there's quite a lot of late night drinking and general carousing. It's to be expected really - it's a bit like a family Christmas (but a family that really likes each other). Often it's the only time people see each other from one year to the next. There's a lot of catching up to do, not to mention all the new people to meet. Pete and I long ago realised that many people went to comic conventions to meet and drink, rather than to fill gaps in their run of Simonson Thor's. The ACCU conference has a similar social buzz. This year, though, was even more boozy than most.
Large-Scale C++ Software Design by John Lakos is one of the early classic books on C++. It was among the first, if not the first, to really address the difficulties that arise when working on large code bases. It provides a variety of solutions to those problems and the trade-offs between them. Much of what it talks about applies beyond C++, but at the time Java and C# were but a twinkle in their corporate sponsors' eye. The book remains relevant, some 11 years after publication, and really has no serious competitor. It is, without a doubt, an important book and if you work with C++ or large systems then you should have read it. It is, however, as dry as dust and I found it very hard work. Canvassing suggests I am not alone.
John Lakos spoke at accu2007. The mental model I had of a tall-ish, quiet, perhaps slightly shy academic were exploded when I realised the scurrying, ebullient, pugnacious even, New Yorker who kept buying drinks was Lakos. Smartly dressed, combative in conversation, fiercely clever, seemingly indefatigable, Lakos was one of the surprise hits of the conference.
As he bought another round at 4am on Saturday, he declared "I love this conference. This is great". Hope he comes again.