Most helpful critical review
Far from great
on January 25, 2002
I thought (and still think) that it is a good idea to write such books. Well written standard specifications (in terms of readability and immediate applicability) are rare, too often they are too dry or "academic" for immediate use in all-day programming practice (best example is the MIME specification, which stretches across several complex RFCs, which constantly reference to each other).
This is where "companion" books like this come in, providing the reader with guideance, with information about real life situations and examples. They point out pitfalls, emphasize parts of the specs that are "more important" than others, or are (or can be) implemented differently in reality.
Alas, in my opinion the book falls short of achieving this goal. It barely does more than picking (sometimes random) parts of the specs and retelling them in the author's own words.
I have yet to find a book about email that doesn't confuse its reader with acronyms. It seems there is an internal competition between authors of these books to squeeze MTA, MUA, MDA and MRA in one sentence as often as possible.
Another example of the author confusing the reader is that the text keeps mentioning gateways to other (proprietary, non-internet) mail systems, and that/how messages must be converted between these different mail systems. While it's OK to mention this *once* in the introduction, this is clearly not the focus of this book, as it is geared towards folks who write software that handles INTERNET MAIL messages, and not gateway implementors (at least this is how I interpret the book's title).
Also, quite often there is bad coverage of the real world. For example, two chapters are devoted to understanding and implementing MIME, including a lenghtly explanation of "interesting" message types like message/external-body and message/partial (both of which I knew of, but have never ever encountered in my work that includes a fair amount of mail
programming). On the other hand, the "multipart/alternative" type (which is very common in today's email world - ever got a message from Amazon.com?) is barely mentioned at the very end of the MIME discussion.
All in all, it's a good idea from O'Reilly to have such a book, but the execution definitely needs polishing.