I kept seeing posts - here, at advogato, and at slashdot -- asking for definitive books. I've seen requests for C, for Unix, for Compiler Theory, but also for stuff like Nanotech, or Evolutionary theory. Anytime someone makes one of these requests, suggestions and pros seem to crawl out of the woodwork with useful suggestions, and with alarming frequency, consensus emerges: it seems some topics just have one or two works that simply must be read. That's where it started.
CanonicalTomes.org is a site I've glued together in the past month, after an initial comment on advogato about creating it garnered me several offers of free server space, if I would put the site together. It's DB-driven (postgres) and perl-written, and is dedicated to the task of cataloguing the definitive tomes of every topic - and every subtopic -- that its users can generate. Users can add books, create new subtopics, and vote on submitted books -- the hope being that with enough eyes, the clear winners in an area -- or perhaps just several compelling candidates -- will rise to the top.
The site in its current state is fully usable, but because it lacks publicity, it only has about 35 books -- enough for us to test with, but not enough to really be a useful resource yet. I'm hoping the k5 community will be the start of something -- it has to start somewhere -- and start checking the site out, and adding their favourites. I'd also love feedback on any part of the site that merits it, especially any new FAQs of note. I'm also putting this up here for discussion because of the data: people have expressed general interest in the data this site will eventually possess -- but I'm not sure what formats would be easiest to work with, and what type of applications people might see for it. I know this sort of feels like shameless self promotion, but I also think canonicaltomes is the kind of site that might appeal to the k5 readership, so here's hoping it generates a little traffic -- and discussion.