well in the case of thomas, the book has been considered the standard for over 40-50 years. In that amount of time however Thomas has actually died and students have changed in their level of, or lack of, preparation. Thus recent editions are completely different from older ones. I.e. it is impossible to recognize that the 9th and the 2nd editions are the same book and in fact they are not even by the same person.
Because the Thomas name was a guarantee of high sales, the publishers just hired other people to completely rewrite the book, and these totally different books were published with Thomas' name tacked on cynically. E.g. well meaning people like the posters in posts 2,5 above are innocently recommending in good faith, books they apparently do not realize have almost no resemblance to the one you are asking about. I heard the same thing on my first job. People who had learned from the original Thomas, always swore it was the best book to adopt, even though they had not read the versions we were adopting. As preparation for my job, I read Thomas' 1st or 2nd edition, then arrived to find we were using a very different edition. Still in those days it did resemble the original in some essential ways.
Now I admit I am often much too hard on decent books, especially when I am frustrated because I am forced to teach from them, but if you are interested in my scathing review of a recent edition you can read it on amazon. The 9th edition, blue cover, lighthouse, however, although totally different from the original Thomas, was very clearly written, and i liked it.
By the 10th edition however, I had completely become disillusioned with it. They are continually dumbed down, to accommodate weaker and weaker students. In the 9th edition a I recall, most proofs of derivative laws were put in the back, but it seemed ok to me to just give a few to get the idea. Then in the 10th as I recall, ALL proofs of these laws were banished to the back of the book. This was the last straw that broke my patience.
So basically the newer versions are giving you less and less information, and challenging you less and less.