On and off, I've been thinking about what I'm going to do with my next book. I mean, I'm planning to write the thing, and to see to it that it is published by a decent publisher. So much is required for it to get the kind of institutional recognition that is necessary for people in a position like mine.
However, I'm looking for something more than that. I'm looking for my book to be read. Getting read is more difficult than getting published, given the avalanche of academic material published each year. I'm hoping for this book to be read, for it to be assigned as a textbook, for it to corrupt the minds of the youth, and for it to shape the field for years to come. (I may as well aim high.)
I have come to the conclusion that one very good way for me to do this is for me to give the book away for free: I plan to post the book chapters, as they're written, online here, to have them indexed by search engines, to get people like you linking to it on your website, glancing through it, giving me feedback, helping me to improve it, etc., and to link to it, making it appear higher in search rankings, and giving other people more ways to stumble onto it, etc.
Now, there are quite a few examples of this in the publishing world already. People give away electronic copies of their book, and despite this more people buy the physical thing published by the publisher. After all, there's something useful in letting the experts print and bind a copy, instead of you dealing with all of that hassle.
Can this be done with academic books? It turns out that it's already been done here too. Allen Hatcher, the topologist at Cornell, has published his Algebraic Topology textbook with Cambridge University Press, and he still offers the whole thing for a free download at his site. This is unbelievably smart, if Allen's aim (as I suspect it is) is to be read by readers. Allen's distribution rights are more restrictive than some public distribution conditions: he, and Cambridge University Press, allow you to print a single copy of the book from the files for personal use, but not to photocopy multiple copies. This is not exactly simple for them to enforce, but you can understand why they make this restriction. The files distributed are completely unencumbered PDF files, which print just as nicely as the Cambridge University Press version of his book. (This, in my view, is much better than the state of play with most e-books which are typographically very poor. That's not too problematic for a novel, but it's impossible for a mathematical text which requires fine typographical control to get the mathematics both correct and legible.) What is the end result? Allen Hatcher is (at the time of writing) Google hit number 4 on algebraic topology, his book is how I learn any algebraic topology I need to know, and the dead tree version for sale at Amazon has a much higher sales rank than my book on substructural logic.
So, I think I will head down this path with The Next Book. Do you have any advice for me? Is it a silly idea or not?