NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/std/c/189

Internet Message Format | 1996-08-06 | 15.6 KB

Path: solon.com!not-for-mail
From: [e-mail] (Scott McMahan - Softbase Systems)
Newsgroups: comp.lang.c.moderated,comp.lang.c,comp.std.c
Subject: Re: Perhaps it's time the C community *did* something about bad books.
Followup-To: comp.lang.c.moderated,comp.lang.c,comp.std.c
Date: 25 Jan 1996 20:41:12 -0600
Organization: Interpath -- Providing Internet access to North Carolina
Sender: [e-mail]
Approved: [e-mail]
Message-ID: <4e9f08$[e-mail]>
References: <4e07lv$[e-mail]> <4e5ooa$[e-mail]>
NNTP-Posting-Host: solutions.solon.com
X-Newsreader: TIN [version 1.2 PL2]

Ray Dunn ([e-mail]) wrote:
: In referenced article, Peter Seebach says...
: >I call your attention to Schildt's "C: The Complete Reference".
: There is a long crit of this book, with many errors listed, at:
: http://bit.csc.lsu.edu/tutorial/ten-commandments/schildt.html

Not exactly. This is actually the review of the ANSI C standard.
The "review" errs on the side of being way too critical, and I've
seen it being debunked in part, but the parts that are debunked
never seem to be posted with the review anywhere. It's one of those
things that will endure forever.

I did not see the post that listed errors in C: TCR. I'm interested
in reading it.

I agree that we should all make Osborne aware of all the errors in
Mr. Schildt's book so that he can make them better in future
editions. I was very disappointed in the 3rd Ed. compared to
the 2nd.

Scott

PS: Here is the review of the 3rd Ed I wrote when I got it last
summer.

C: The Complete Reference, Third Edition
By Herbert Schildt, published by Osborne/McGraw-Hill
ISBN 0-07-882101-0

A review by Scott McMahan, July-August 1995

---------------------------------------------------------------------------

I finally got my copy of C: The Complete Reference -- not a review copy,
as I've never been able to make contact with any Osborne press people.
I don't know if they even send out review copies, but if they do I've
never been able to get my foot in the door.  I thought I'd give the book
a once over anyway. It took me some time to complete this review, for
various reasons, so it isn't as timely as I thought. On the other hand,
the third edition seems to have been released in extreme secrecy since I
learned about it only though a McGraw-Hill book-of-the-month club ad.

Way back when, I learned C using Schildt's Teach Yourself C, and have
always had a soft spot for his writing.  He is a capable author who has
the ability to simplify complex topics enough for people who have not
had exposure to them before.  I'm certainly not suggesting his books can
replace Sedgewick's Algorithms In C or K&R2 (both of which I consult
regularly), but they give a kinder, gentler approach to topics that
prepares you for the more precise and "heavy" books you'll have to read
later on.  If anyone is willing to give Herbert Schildt a fair shake, it
would be me, since I have benefitted from his writing in the past and
have read a significant number of his books.  I've even defended some of
the wilder acusations about Schildt on the net, but after the third
edition I don't think I'll bother.  I intend to send a copy of this
review to him in care of his consulting office in order to hopefully
make the fourth edition much better than the third.

Just picking up the book, one obvious thing is the print is smaller.
Holding it up to an earlier book like Turbo C/C++:  The Complete
Reference 1st edition makes the difference startling.  Paper costs have
gone up tremendously in the past few years, and shrinking the type books
and other things are set in is very common. It's unfortunate for the
eyes.

The page layout has been revamped, with little improvement. Big grey
bands set off code from the text, which is unnecessary and makes the
chapters on functions (each of which comes with an example program) very
chaotic looking. The bands overpower the headers, which is what you're
skimming for.  Compare the stylized 'C' with the logo of my C And C++
Review -- wavy pixilated designs with the foreground blurred into the
background.  I'm sure this is nothing but a coincidence.

What's up in the third edition?

Unfortunately, Schildt has chosen not to correct the void main(void)
anachronism. From what I understand, the usage which is clearly and
flagrantly contrary to the ANSI standard is a throwback to pre-ANSI DOS
C compilers' rather strange ideas on returning values to the calling
environment. (I don't know why this is even a problem, as returning a
value in register AX is incredibly commonplace to any DOS program. Most
assembly language utilities you run across return a value to DOS use
register AX for their processing to kill two birds in one stone: #1 not
having to clean up the register, and #2 returning the value as a
side-effect of the computation.)

To make this issue even more bizarre, Schild says on page 163: "The
main() function returns an integer to the calling process, which is
generally the operating system." The part is absolutely true (the part
about the operating system isn't -- usually programs are started by some
type of command interpreter doing an exec() or spawn() call, and the
command interpreter gives some meaning to the return value). Then, from
this very stable ground of ANSI behavior, Schildt takes a flying leap
off a cliff and says in the next paragraph: "You may also declare main()
as void if it does not return a value". Since he has just said in the
preceeding paragraph: "If main() does not return a value, the value
passed to the calling process is technically undefined. In practice,
most C compilers automatically return 0, but do not rely on this
behavior if portability is a concern." -- it makes me wonder why he
doesn't just explicitly return a value, instead of hoping for the best
from undefined behavior?  (And remember what undefined means: The C
compiler can do ANYTHING it wants, including what the GNU compiler did
to protest the inclusion of #pragma in the standard, start a game of
rogue.) A void-declared main is undefined. An integer-declared main is
defined, and just returns garbage. Garbage is better than undefined
behavior.

The silliest thing is "main()" is a lot easier to write than "void
main(void)", which is equivalent to using stilted, awkward phraseology
in English. As a writer, Schildt should at least have his inner ear
tuned to the easiest, best-flowing way to phrase things. Additional
silliness comes from the fact that some of the material is 5-10 years
old, and some of it is brand new. Thus some parts of the book use void
main(void) and some don't. The little C interpreter, for example, does
not parse this syntax. (It was, in fact, written before ANSI C, and its
syntax doesn't include the ANSI keywords -- as much as Schildt likes his
void affectation, this is surprising. Adding the "void" keyword to an
interpreter would be little more difficult than ignoring the word
"void".) Another example is that all the Windows' functions are declared
as "int WINAPI WinMain", which makes little sense as the return value of
a WinMain is as clearly defined as the return value of an ANSI C
function.  If one definition of program behavior is to be ignored, why
aren't all? This is inconsistent.  It should be void WINAPI WinMain.
This is as undefined as the ANSI C definition of main()'s declaration.

The book is still very much PC based.  It isn't a true ANSI C reference
like the book would like to mislead you with the front and back cover
copy. This is unfortunate, because the truth-in-advertising issue gives
Schildt a lot of bad press. If they just said "for PC programmers" on
the cover, they'd eliminate a lot of the vitriol the book receives. A
lot of stuff is outright PC based (as described below) and would work on
nothing but a PC.  A lot of the other stuff is based on the assumption
that you're on a PC (such as the size of a word is a 16 bit half-word).
It would work, but differently, in other environments. This is the
dangerous stuff. Instead of calling it a Complete Reference, it would be
a lot more precise to say that the book is a superset of ANSI C for PC
programmers.  (I think "complete" is a joke anyway, like the "complete"
Internet reference. How can anything written about the Internet be
complete...!?!)

In addition, Schildt discusses only the Microsoft and Borland compilers!
No mention of Mix, GNU, Watcom, or any others.  This isn't a complete PC
reference, either. If you're going to talk about PC compilers, and be
complete, you need to cover compilers as popular as Mix's Power C. I
understand a book like Turbo C/C++: The Complete Reference covering one
specific compiler, but when the door is opened to all PC C compilers,
coverage of the major players is necessary.

Topics of interest to PC programmers include:

    The segmented 80x86 memory model, which is a non-issue now that
    Windows 95 allows you to write flat memory model programs (the
    same Windows 95 about which a chapter is included in this very
    book, see below);

    Microsoft C/C++ console graphics functions (which are for some
    inexplicable reason claimed to have been deleted, in the
    introduction, but are still present in the text! Chapter 17 is
    nothing but the Microsoft C/C++ 7 DOS-mode graphics chapter in
    the MS manual, warmed over);

    DOS assembly language, both compiler output (for efficiency
    purposes, ironically he's using an old CISC chip for his
    efficiency material) and for interfacing C and asm.

That's an interesting cross-section of DOS. Lots of bizarre looking
functions like "_intdos" and "_intdosx" appear, which some compilers
allow you to use. (Come on, if you're going to call DOS interrupts, why
not use embedded assembly language? If you're going to be non-standard,
using the nonstandard asm keyword is a lot more explicit! The asm
keyword is more widespread, at least, and the assembly language screams
"I'm a non-standard module".)

On p.  8, Schildt claims spreadsheets are systems programs.  I never
really considered that to be true, but then again OS/2 Warp does, in
fact, ship with a spreadsheet as part of the basic OS package.  I guess
times change.  Once, spreadsheets were applications that you had to (and
still have to if you want to do anything serious) buy on your own.  As
you can guess, this is probably a result of the years of cutting and
pasting done to the book.

What's new?  The third edition is surprisingly unchanged. Instead of
taking the chance to reflect on which chapters he has written in his
almost 10 years of writing on C and picking the ones that would be most
useful to an ANSI C programmer, Schildt has instead placed chapters that
have little to do with a complete reference, including a tacky
self-promotion.

A C interpreter:  this is from the out of print book Born To Code In C,
although I am under the impression that the material in Born was also
reused in The Art Of C and The Craft Of C.

It is hillarious to contrast the wording between Born To Code In C,
where the C interpreter is in the first chapter, to the third edition,
where it is in the last chapter.  Either one by itself reads fine, but
when you contrast them side-by-side, the results of the cut & paste job
are side-splitting as Schildt's enthusiasm and sincerity are shattered
when you see how he changed a couple of lines.

The good thing, though, is that the interpreter source code is readily
available.  It's not a bad program, although I haven't typed it all in.
It looks like it does a decent job. A lot of people want to have a C
interpreter. With Born To Code In C out of print, this makes it availble
to everyone.

(This is another big beef I have with the book: the blasted thing is
$35, you think that they could have put a disk with the source code in
with the book! I've seen books a lot cheaper than this with disks.)

It's nice to have the interpreter and the recursive descent parser in
the same book.  (Schildt's continued insistence -- probably more because
he didn't read the chapter on the parser when he cut & paste it than
anything intentional -- that recursive descent parsing is somehow
mysterious or secret is not really in keeping with the times.  Most
college CS courses on recursion cover them.  The only people to whom
recursive descent parsing would be mysterious are users who are not
sophisticated enough to understand algorithms in general.)  Having them
separated makes referencing one while reading the other a pain,
especially when one book is at home and the other at work.

(Perhaps a coincidence, perhaps not:  the section on ideas to extend the
interpreter has as one suggestion adding an integrated editor: exactly
what Quincy, the interpreter Dr. Dobbs' Al Stevens wrote.)

I almost took the book back and demanded a refund when I saw this. A
chapter on Windows 95 is nothing but a shameless, tacky ad for Schildt's
new Windows 95 Programming book.  They should have put this chapter at
the back of the book and clearly labeled it as an ad.  Windows 95
programming has no place in a generic ANSI C book.  Schildt just took a
chapter out of his Windows 95 book and pasted it in here. It is not ANSI
C, it is not even particularly C, it has nothing at all to do with a
"complete" reference since the skeleton doesn't do anything, and
all-around is just plain bad. The bad feelings generated by this
sickening ad will more than destroy any small sales increase M-H gets.

What isn't new is the sad part -- with the disgusting ad and the C
interpreter (which although useful is certainly not anything you'd
expect in a C reference), Schildt did not include the chapters from
Advanced C on statistics and random number generation.  For an ANSI C
programmer, this material would have been much more useful.  People are
asking for random number generators all the time on comp.lang.c, and the
older Advanced C book has several. Advanced C, which is now out of
print, has a lot of information that would be useful to the general ANSI
C programmer. I asked Schildt to consider including these chapters, to
which he responded with a positive message that he'd consider it, and
yet greed and advertising won out over the usefulness the book would
have had to people. I see requests for random number generators on
comp.lang.c frequently. It is material people have an interest in and
need to know regardless of what C implementation is used.

There's also not much on hashing, one of the fundamental things you need
to do in C.  I had to write a hash very recently, so I though I'd see
what Schildt said about it.  Very little, as it turned out, and only as
an aside in a chapter on sparse arrays.  The bad part is, he doesn't
explain hashing much at all and only has an example of closed hashing.
(My own table, for example, was open hashed.  I had a situation where
open hashing made the most sense.  Who names these things, anyway?  The
type of hashing with a finite number of slots in the table is "open" and
not "closed"?)

Unless Schildt starts trying to find what people want and starts
remembering why he wrote C:  The Complete Reference in the first place,
he and Osborne are going to have a hard time selling books with ads for
other books instead of content and using Schildt's name.  Schildt is
already known as a schmuck among real C programmers, deservedly or not,
and pretty soon everyone will be calling him the "recognized authority"
on self-promotion and hot air. Either Schildt or McGraw-Hill appears to
be damaging Schildt's already very shakey reputation beyond repair for
small short-term gain. I don't know why, is Osborne in some kind of
financial trouble? They should be trying to build a reputation for
quality.

Unfortunately for me, I will probably never order another edition of
this book, unless the concerns raised here are directly addressed.  I'm
very disenchanted with this book, and will look elsewhere for my
reference material.