This is a problem in calculus. I tell people about the Nth edition of <standard text>, which is about $200 new. I have a few with (N-1)st edition, and a few with the Early Transcendentals version, and a few with the (N-1)st edition of the Early Transcendentals version. I point out to them that I'm not responsible for the differences, although they should know this.
One problem is that the library doesn't actually have a copy to put on reserve, and I only have my desk copy. Another problem that I have is that the next edition (usually no more than four or five years after the introduction of the last) has only a few cosmetic changes from the current edition, just enough to ensure that page numbers, section numbers, and homework problems don't match up.
There are open-source equivalents to many textbooks out there, and I might consider going with one of those except that I think it will be a long time before any open-source text gets widespread acceptance. I believe that a department would be deeply concerned with a textbook committee that recommended an open-source text, for many reasons.
There's an old saying in the computer world: "Nobody ever got fired for buying IBM." I think this is the same with the traditional standard calculus texts like Thomas, Stewart, and so forth. But that's why those texts cost over $200 new, and come out in new editions every few years with just cosmetic changes; because nobody has the guts (or the time) to create and champion an open-source alternative.
It certainly won't work in departments where instructors use the same book, but last year I was able to "trade" an extra copy of the text a colleague was using (and I wasn't) for his extra copy of the one I was using so I could put one on reserve.
Then again, I'm not sure students read the book if they only have access to the reserve copy for 2 hrs at a time at the library (our only reserve option). But that's an issue for a different thread....