When I receive any new issue of the CUJ I usually open it at an arbitrary page. In the March issue the first thing I read was Bob Moore's complaint about your review of Anthony Porter's book The Best C/C++ Tips Ever. Let me assure you that I was more than pleased with your review and especially glad for the unequivocal words with which you started it. So I wish I had sent you this e-mail sooner.

Being a reader of the CUJ for many years (in fact since long before the "U" appeared in its name), I think I know the moderate tone you usually chose when you verbalize criticism. I consider the idea there could be any connection between the two facts that you published some books at Prentice-Hall and Porter's book is published by McGraw-Hill to be silly, if not impertinent.

I learned programming in C around 1981 (using K&R-I), and as I'm teaching C for more than ten years now, I have some professional interest in books about this topic. So, every few months I spend some hours in a well-sorted book shop to see what's new. In the early 1980s there were so few new books that I bought practically all I could find and read them cover to cover, if only to decide if I could recommend them to the participants of my C courses.

During the late 1980s I had to give up that habit, not only because it would have been too expensive to buy all that stuff but also because of lack to time to read it, even if only partially. But I continued my occasional 2-3 hour visits in book stores to glance through new books. The few I bought were mostly to get some up-to-date things on my book shelf (like K&R-II, H&S-II+III [Harbison and Steele — pjp] ...) while ANSI C approached its final shape. My general feelings were that anything that could be written about C was already in some book, so why do people still write so many more books?

In the early 1990s I followed the Usenet-discussions in comp. lang.c, and occasionally made contributions myself. By the helpful and friendly people there (Karl Heuer, Chris Torek, Henry Spencer, to name only a few I immediately remember from these days) my expertise in C got its final polish. I especially learned to rate the FAQ of comp. lang.c very high because of its sound mixture of low-and and high-level stuff. Steve Summit does a really admirable job in explaining things in a terse, yet not over-simplified way.

Given that background, I'm surely not the "typical reader" to which new books on C are aimed, but my general impression if I visit a book store today is that most of the stuff is at best unneccessary, if not — in direct comparison to some of the "classical" books " — bad or even "dangerous" (as you clearly expressed it in your review). I can skim through almost any new book and it will typically be less than five minutes until I find something that wouldn't have survived if I had been in charge as technical editor. (In comparison, it's extremely hard to find such things in K&R-II, though I'm aware of some weakness on pp. 119-120.)

It's not that all the things I have to criticize are plain wrong. (Some are. I seem to have an especially sensitive eye for syntax errors in examples, which often are obviously not included from the source files but re-typed.) An example of what I typically spot are system dependencies not marked as such, say discussions on the behavior of character input with getchar, which may be true for the system the author uses but cannot be generalized. While the novice reader is given the impression the author knows a lot about C, to me it more reveals the author's limited experience and exposure to a number of different implementations and/or operating systems.

As I know that one might be tempted to use on-line help and silently get lured into implementation-specific features, my "standard reference" (pun intended:-)) is the small book you wrote with Jim Brodie. (BTW: It may be misleading to some people that you choose Microsoft Press as publisher, because this book is as close to ANSI C and as far away from any MS-specific stuff as I can imagine.) I would urge every author of any C book which claims to deal with ANSI/ISO C to use this booklet as reference — or at least to cross-check his or her examples with it.

Failing more and more to find recommendable new books on C I finally began to doubt on my ability to accept other views than mine. I mean, could it be that the problem was more with myself than with the books? Maybe it's the growing indifference to other opinions and styles that some elder people develop? Well, I do not really feel old (I'm forty now), but maybe those are the first symptoms. So your review of The Best C/C++ Tips Ever was kind of a relief to me. I remembered having had a look into that book too and put it back onto the shelf very quickly. (To be honest, because I already disliked its "promising" title, my acceptance level was set a bit higher than usual — and the more the book failed to reach it.)

If I may express a wish for further book reviews in CUJ: Please continue to be explicit with criticism of bad books. It was reported to me that someone once wrote in comp. lang.c, "All the books on C you ever need are K&R, H&S, and the FAQ of this group." Though this sounds provocative (probably it was meant to be) and there certainly are other recommendable and useful books, I fear most of the new stuff is unnecessary and some is hardly worth the paper it is printed on.

I suppose that not only the community of C users grows but also the number of authors who try to place their books as best they can. Currently the latter growth (books) seem to exceed the former, so many publishers and authors will have a hard time to get their share. Some obviously try to get attention by putting "marketing hype" in the title (best tips ever). Others simply update old books so that the buzz-word "C++" can show up in the title of the new edition — and if you open the book you'll find mostly C stuff with a few C++ "add-ons," typically not "good" ones in the sense of supplying an idea of what OOP means. Sad as this is, I think we — the "oldies" — have some obligation to warn the newcomers about it. I appreciate that you did this in the January issue of the CUJ.

Best regards and keep up your good work,

Martin Weitzel

Greetings:

Re: Letters to the Editor, March 1995, Concerning The Best C/C++ Tips Ever

Excoriatus, exculpatus.

I read with delight your reply to the mindless rantings of Bob Moore. The article at issue rendered a service to C/C++ users. That is what this is all about, isn't it?

If the ruffled Mr. Moore wants to be helpful in the same way he could name names regarding the alleged buggy software. If he has the courage to identify it I'm sure you'd publish his article.

Perhaps leaving unedited his thorny prose, once again.

Sincerely,

K.B. Williams
Stillwater, Oklahoma

Okay, here goes. See letter following. — pjp

DR. P J. Plauger:

Thanks for your chastisement in the March issue of C/C++ Users Journal as a response to the rudeness of my letter regarding your review of a programming book. Frankly, it's almost an honor to have been mentioned in any context by one of the computer giants of this age. Had you been a little more perceptive you might have characterized me as not just as a mud slinger, but an exceedingly ugly, short, fat, bald, ignorant mud slinging little troll who lives under a bridge near old Fort Tombechbie and rides a splay footed donkey to work. The donkey would probably have been insulted.

Most people who have spent time in residence on planet earth and been involved in some form of programming would probably parse roughly the same level of significance and viability in the following two sequences of tokens: "(IBM) PC compatible" and K&R (&P)." While the latter sequence seemingly is just arriving, it exasperates me to no end that the former sequence is losing impact.

I lament the passing of the IBM business ethic that we've all enjoyed for over a decade. With entrepreneurs coast to coast in this country and around the globe cloning every known item of IBM hardware, most of them also cloned the IBM business ethic, which was perhaps the best thing ever cloned. In my perception, at least, that ethic can be interpreted as meaning, "If a customer buys our products and adheres to the written terms of our warranties, then that customer will be entitled to all and complete benefits to be derived from IBMness." Even trolls.