This is a good idea, and one whose implementation has already been begun, to some extent, in places like Wikibooks.<p>OTOH, if we're really looking to write good course texts, then let's remember what a text is all about. It is, of course, first and foremost, something to read to get a thorough introduction to a topic. But it is also a reference. From an instructor's point of view, it is often a source of homework exercises.<p>Also (here's a tough one) it generally functions as a partial course outline. That means that starting from the beginning and plowing through, should be a reasonable order of topics for a course. In addition, the better texts usually have a section listing dependencies among chapters and giving alternate course outlines that the text can be used with.<p>The trickiest issue, I think, is the question of authority. A textbook functions as an authority in a course, second only to the instructor. We're supposed to be able to rely on it. This is the same issue confronted, and basically sidestepped, by Wikipedia, which was successfully able to turn the focus from authoritativeness to the related issue of quality of information. Since Wikipedia clearly shines in the latter, we don't worry so much about the former.<p>Can we do this same sidestep for texts? I don't think so. Universities, in particular, have accrediting bodies, which will not look favorably on something with standards along the lines of Wikipedia, being presented as authoritative. Thus, your idea of the effort being peer-reviewed. A good idea, but how to implement it?<p>Lastly:<p>> Somehow, I don’t see many scientists relying on royalties for a non-trivial portion of their income.<p>Certainly not. I believe some do, but, if they can't compete with <i>free</i>, then I figure that's their problem. I wouldn't worry about it.