Title: Editorial

I have recently had a couple of authors getting very upset because I have written a negative or lukewarm review of one of their books. Before I go on, let me remind readers that I have always stated that those who disagree with a review should write an alternative that under certain constraints (not written by the author, his relations, friends or publicists etc.) it will be published. Sometimes I, or other reviewers, make factual errors that we are always happy to correct (while feeling embarrassed that we made them in the first place). However, when it comes to assessment of a book; that is always a reviewers personal opinion. The reviews we publish never (well very rarely) have an anonymous or pseudonymous by-line.

We need to distinguish between situations a single good item is sufficient from those where multiple instances are desirable. If you are a fan of detective stories, you certainly want more than one. You will even tolerate the reworking of a standard plot (well you better, because there are not that many available). While wishing that every book were a literary masterpiece for its genre, you will happily buy and read lesser examples. The same holds true of all varieties of fiction but as we move away from fiction our tolerance, as readers, diminishes. Even if you are a devotee of Shakespeare or Winston Churchill there is a limit to how many biographies of them you would read. You would not want to read a new one even if it were quite good unless it was fairly original.

This is even more true of performance art, you would not wish to spend time and money on a mediocre performance of a piece of music, a ballet or a play. I know there are a few people who get stuck in a groove and, for instance, go and see the Mousetrap once a month, or once a year. But even these will go to the authentic version.

So let me now consider technical books. You will only learn C, C++, etc. once in your life. Having done so, you will no longer be interested in learning the same language again. The book you use will either have served you well or not. The process of learning is expensive, in terms of your time if nothing else. There is certainly room for more than one introductory book per language because we do not all learn in the same way, nor do we have the same prior experience. A book on C++ aimed at the newcomer to programming should be different to that aimed at someone who already programs fluently in C, and both should be different to one aimed at Smalltalk programmers.

One particular author claims that his books that are very specifically aimed at first time programmers have been helpful to those already drawing a salary for writing C++ code. I think he may be right, but that only highlights the terrible state of ignorance of not a few who earn a living by purporting to write C++. It says nothing about the suitability of his work for the target readership. The same author defends the contents of his books on the grounds that little has changed in the specification of C++ since 1995. Even assuming that his books were of good quality then, that says nothing about their quality today because the way C++ is written has radically changed over the last six years. The traditional approaches to introducing C++ to novices are certainly very demanding of the novices understanding and dedication. The technology now available means that we have cleaner, simpler ways of getting novices started than those current six years ago.

The publishing industry would do a great service to all by restraining its 'me too' feeding frenzy whenever some new topic gets hyped. While a poorly written novel may still have something to offer, most publishers decline to publish fiction that is written in grammatically incorrect English (German, Japanese etc.). The place for such work is on the Internet where those with more time than sense can boost the author's ego. Of course, even the best of fiction contains the odd error, Even worse they sometimes contain factual errors (towns in the wrong country etc.) which we excuse because they are not germane to the story, just irritating. We should apply the same standards to technical works, and when they are about a computer language the source code should be correct. If it is not, the book should be rejected without further argument. We can be more tolerant if the book is not directly on the language but about some other aspect of programming.

Publishers ask for opinions on books offered for publication. One of the pieces of information they seek to ascertain is whether the proposed book has direct competitors. Too often the answer seems to be ignored when there is an excellent competitor but from a different publisher. In fact, I would suggest that quite a few books come dangerously close to plagiarism.

Now an idea that crosses my mind is for ACCU to offer a service to publishers whereby book proposals are placed on our website and everyone is free to comment on them. Perhaps some refinement of this idea would help to reduce the number of mediocre books cluttering our bookshops. Unlike services, there is really no value to a technical book that is second best, if it does not have a distinctive value of its own then it is a waste of space. That does not mean that we only need one book for novices per programming language, but it surely means that we do not need dozens even if they were all technically correct.