There is a 1974 (0385067100) and a 1984 (0713442107) version of this book. The 1984 version is advertised as "revised." Is the content of the book revised as far as errors and theory is concerned, or is it only a mere conversion from DN to AN, or both?
Just a conversion to Algebraic. Not too many errors in it, unlike the algebraic version of Rook Endings by Levenfish & Smyslov. A disaster there, nothing but incorrect diagrams, misplaced analysis, and typos all over the book.
Hi, I'm pretty sure the material in the two versions is identical. I only had the 1980s algebraic one, but it reprints Keres's preface from 1974 which suggests an unaltered reissue, and Keres died in 1975, so he can't have done any updating himself. It's therefore purely down to which notation you like (unless you're into condition/collectability considerations).
The book does have a well deserved reputation as one of THE chess books to own and study.
Thoroughly agree with the comments on the 1989 algebraic edition of Levenfish and Smyslov, I've never known anything like it for the sheer number of misprints.
You're right, he did die in '75 - right after the book got published! He writes and publishes one of the best chess books on the subject and, plop, he's dead. I wonder what people at the time said about this?
I think I'm going to go for the '74 version in DN because it is a more accurate reflection of the period the book came out in.