C: The Complete Reference (Osborne Complete Reference Series) (Paperback)

Comparing it against other C reference books I now see the deficiencies. Schild's book tries to be all things to all men:

oOo Tutorial: A very chatty manner as it leads you through the workings of each area, I like the description of "make", albeit brief.

oOo Reference manual: It covers most of the C language, but there is incomplete coverage of the C language (eg errno, and bzero).

oOo Algorithm book: It describes the use of Artificial Intelligence, but this is such a huge area it only wets the appetite. A reader attracted by these areas would better off buying a book which concentrates on these areas specifically and in greater depth.

I was disappointed to discover the third edition of this book had the C++ overview removed. The new section, "a C interpreter", which I suspect is to pad out the book after removing the useful chapters from previous editions is practically useless. It would have been better to write something similar to Steel's coverage of writing 'clean C' where a C program runs in a C++ environment to distinguish the conflict areas between C and C++.

Another failing of Schild's book is the all too brief coverage of some areas summarised with the cop out "consult your manufacturer's manual for details".

In the format of a good reference book, Schild lays down many parts of the C language under function headings such as "signal". However, towards the end of each description is a section entitled "see related functions" with an incomplete list of references; for example under "signal" it is "raise", but missing "ssignal, psignal, gsignal". Instead of spreading descriptions around like this it would be easier to read if related functions were grouped under a similar heading, so the reader isn't distracted searching for them.

Schild's book has moved from discussing C in a DOS environment in his first edition to a Windows environment in the third edition. Although Schild does cover UNIX, the platform where C originated, he has skimmed the surface like a pebble across a pond.

In conclusion Schild's book is suitable as a tutorial for beginners but cannot be classed as a reference manual. I would recommend the following books instead:

Reference: C : A Reference Manual, Samuel P. Harbison, Guy L. Steele (September 1994), Prentice Hall; ISBN: 0133262243

Tutorial: C : A modern Approach, K.N. King, (April 1996), W.W. Norton & Company; ISBN: 0393969452

Algorithm: Algorithms in C : Fundamentals, Data Structures, Sorting, Searching Robert Sedgewick, 3rd edition (October 1997) Addison-Wesley Pub Co; ISBN: 0201314525

The Algorithm Design Manual, Steven S. Skiena, Steve Skiena (November 1997), Springer Verlag; ISBN: 0387948600

The reviewers who gave this book a 10 have been sadly duped and don't even know it. These people owe it to themselves to get the Harbison and Steele reference manual, and read the Usenet FAQ about the C language, and to read Peter Seebach's partial review of Schildt's book whose URL has been mentioned at least twice in some other reviews here. Also, anyone who doubts the negative reviews should write an article to comp.lang.c with the subject ``Schildt'' and a blank body. That alone should incite a frivolous anti-Schildt flame thread that will last a good two weeks.

The Usenet culture has even coined a new epithet which is applied to a horribly wrong assertion (about a programming language) which is nonchalantly presented as fact: that term is ``bullschildt''. This will no doubt end up in the jargon file one day, thus immortalizing Herb's family name. That just goes to show you that you should be careful about what you publish!

I give this book a 2 only because I believe that the ranking of 1 should be reserved for works like _Learn C in 21 Days_, _C For Dummies_ or others of t! he same trash bin calibre.