The Concise SGML Companion

Sgml

The SGML Handbook

Format: Hardcover
Author: Charles F. Goldfarb
ReleaseDate: 28 February, 1991
Publisher: Oxford University Press, USA
Rating:

Necessary for SGML; useless for XML
Given how broad and confusing the SGML standard is, it's not surprising that this book on it is equally opaque -- this is, in my experience, the worst-written technical book I've ever seen that is not actually inaccurate. This book is, regrettably, the one authoritative book on the SGML standard. But if you're doing serious SGML development, you have no choice but to get this book and to spent forever trying to make sense of it.

But beware: if you're doing just XML, and if you think "well, since XML is a form of SGML, I might as well get the SGML standard", don't do it! XML is all you need to know, then just look at the XML standard, at . . . and maybe also get a book specifically about XML. I happen to like Eckstein and Casabianca's /XML Pocket Reference/, partly because it's less than one-tenth the price of the SGML standard, and a hundred times more useful!.

Required SGML Reading IF
If you are a beginner, or just playing with SGML, this book isn't for you. you are planning on really getting into the world of SGML. The book does contain the entire ISO 8879 standard and is extensively cross referenced. After five years using it, I still find it easy to get lost in the references. If you want the final, definative word from the MAN who wrote the standard, this is it.

Turgid, obscure, confusing; but essential for advanced SGML.

However, SGML is so far the only reasonably universal and standard way of marking up text, and this is the only comprehensive treatment of it, including all the peculiar little bits that you probably should never use. This is a particularly badly written book on a particularly badly designed and written standard, SGML. The book includes the full text of the ISO standard as well as cross-references and annotation.

The book, like the standard, uses terminology and notation which are not standard in the rest of computer science. The tutorial material is weak. The book design is ugly and hard to read.

Yet SGML, bad as it is, is an important and useful standard, and this is a comprehensive reference for it. Let us hope that both the standard and the book will be improved radically in the future.