Accelerated C++ book outdated?

[Andrew Koenig]

Where can I read about this? What new things are included?

And more importantly: Does any of these changes affect the truthness of
the stuff in AC++?

You can find the complete list of changes here:

www.acceleratedcpp.com/authors/koenig/c++std/revisions.pdf

To my knowledge, the only effect that these changes have on AC++ is to
transform some previously false statements into true ones That is, the
book anticipated some of the changes.

 

[David Rasmussen]

I would not call that advice lousy. In fact, the book that taught me C++ was
the second edition of that book - and the third edition is one of the best
C++ books, I've read ever. I find it gives an excellent introduction to C++,
including historical background and paradigms.
Of course, Koenigs book also rates to be first class (based on review and
the extracts, I've read).

The advice is lousy because even if TC++PL is an excellent book, AC++ is
just much more excellent when it comes to _learning_ C++. I am not
saying that you can't learn C++ from TC++PL. A lot of people did,
including you. I am just saying that as good as it might be for that
purpose, AC++ is much better. It is not as complete as TC++PL, and not
as good as a reference. But it is much more to the point when it comes
to teaching good habits and avoiding pitfalls.

I have both books, and they're equally excellent in their niche:
teaching and reference.

I recommend you read and absorb AC++ first, then TC++PL. "The Design And
Evolution Of C++" is also a very exciting read.

/David

/David

 

[Andrew Koenig]

I'm sure that Markus Dreyer's University Librarian
would appreciate a little guidance here.
She doesn't want to buy an expensive new textbook
that may sit neglected on her shelf after Andrew and Barbara
publish their latest, more up-to-date edition.
Can you tell her whether or not she should expect new editions
of "The C++ Programming Language" or "Accelerated C++"?

Barbara and I are not working on any substantial revisions to "Accelerated
C++." When mistakes are brought to our attention, we shall, as we have in
the past, correct those mistakes in subsequent printings.

My most recent publishing project doesn't actually include my name anywhere,
but I had my hand in it anyway:

http://www.amazon.com/exec/obidos/tg/detail/-/0470846747
http://www.amazon.co.uk/exec/obidos/ASIN/0470846747

Barbara and I are both working on other books; I hope you'll hear about them
later this year.

 

[jeffc]

I would not call that advice lousy. In fact, the book that taught me C++ was
the second edition of that book - and the third edition is one of the best
C++ books, I've read ever. I find it gives an excellent introduction to C++,
including historical background and paradigms.
Of course, Koenigs book also rates to be first class (based on review and
the extracts, I've read).

As Mr. Stroustrup himself just acknowledged, both books are good "for their
purposes". TCPL and Accelerated C++ are different books and I think most
programmers should have both, but a reference by definition is not the best
book with which to "learn C++ quickly", which is what the OP was
specifically looking for. The books themselves state their intended
purposes: Accelerated C++ is to help you in writing C++ programs quickly.
It even refers specifically to TCPL for those who want "more". TCPL is
broader and more detailed to help those who want to achieve "mastery". I'm
happy it worked for you, but I disagree with your advice too.

 

[jeffc]

I recommend you read and absorb AC++ first, then TC++PL. "The Design And
Evolution Of C++" is also a very exciting read.

Yes but make sure you keep it under your mattress so no one discovers
it......

 

[David Rasmussen]

Yes but make sure you keep it under your mattress so no one discovers
it......

LOL...

/David

 

[Jerry Coffin]

[ ... ]

In view of the fact that a new ANSI/ISO C++ standard has been drafted,
*all* of these texts are "outdated".

The basic idea here is incorrect. The standard has been updated, but
only minimally so, and most of the changes are merely to make the
standard actually require what was intended in the first place.

A book would only be rendered obsolete if some of the information it
gave became incorrect or out of date, and the recent changes are
unlikely to do that with most books -- the most they'd do is render
obsolete a footnote in a few that says something like "technically, the
memory of a vector isn't _required_ to be contiguous, but it always is
anyway.", because now it IS required to be contiguous. That, right
there, is probably the single most visible change in the standard, which
gives an indication of just how subtle most of the changes really were
(and gives an indication of how good a job the committee did in the
first place, getting that large and complex a document so accurate that
even after years of intense study, only such minor defects have been
found).

 

[Attila Feher]

I suggested our Librarian for the University Library to buy

Accelerated C++. Practical Programming by Example.
by Andrew Koenig and Barbara E. Moo
http://www.acceleratedcpp.com/

but she replied saying it might be outdated, since it was published in
2000, and therefore not worth buying. I don't have the book, I just
saw a lot of recommendations for it on the web. Do you think it's
outdated? Is a new version planned? What is the best current book to
learn C++ quickly given you already have programming experience in
another prog. language?

It is not outdated. You need to find a politically correct way to tell you
secretary to make decisions in her competence area only. I am sure that
Accelerated C++ is not outdated a tiny bit, its content is A OK, even with
the 2003 revision of the standard. Which was mainly a "bug fix" revision,
making original intent obvious.

 

[Attila Feher]

What are the changes?

There are no changes, only "bug fixes", making the original intent clear.
Andrew Koenig has an unofficial list of changes on his web (search
comp.std.c++ for the announcement) or you can buy the PDF version of the
2003 standard at the ANSI store for 18USD.

 

[Attila Feher]

That's probably correct.

Completely incorrect. It is a myth, that the 2003 standard is a different
C++. I would say it is basically updates for clarification.

Andrew may want to answer this.


"The C++ Programming Language: Third Edition" by Bjarne Stroustrup.

Quickly is the keyword. And I believe that (knowing the average ability of
most novice programmers to read and study non-obvious text) while the above
book is very good, I would not call it a quick introductory book of the
language. It is a rather heavyweight material, although it does not bite
your ear off.

 

[Attila Feher]

In view of the fact that a new ANSI/ISO C++ standard has been drafted,
*all* of these texts are "outdated".

False. There is no C++ standard out, which you could call *new*. There is
a Technical Corrigendum added to the existing standard to clarify its
original content.

 

[Attila Feher]

Peter Koch Larsen wrote:
[SNIP]

I would not call that advice lousy. In fact, the book that taught me
C++ was the second edition of that book - and the third edition is
one of the best C++ books, I've read ever. I find it gives an
excellent introduction to C++, including historical background and
paradigms.
Of course, Koenigs book also rates to be first class (based on review
and the extracts, I've read).

I guess if you read the two books you will agree that AccCpp is much closer
to a quick and painless introduction than TC++PL. This does not judge
any of the books negatively, they are just different. There are people who
have no problem using TC++PL and get good results,but IMHO the average would
just give up or suffer. And remember: we talk absolute beginners, who have
learnt Pascal or something but did not do real work yet.

BTW Francis Glassborow did an interesting beginners book, for people not
knowing programming at all(!), and uses C++ as a language. ;-)

 

[Attila Feher]

Your librarian is a C++ expert? Wow. Is she cute? ;-)

I would answer no to the previous question. But if she is cute I can help
about that.

 

[Attila Feher]

It isn't out of date.

In fact, there are some statements (those having to do with
value-initialization) in the book that were not quite true of the
C++1998 standard, but are correct for the C++2003 standard. Barbara
and I made a guess that the standard was going to be revised and
wrote those parts of the book so that they referred to the revision.
We also took pains to be sure that all of the examples in the book
would work with both the C++1998 and C++2003 standards.

To back up what Andrew Koenig has just said: I have thaugh 20 beginners
using AccCpp at my workplace some time ago. The book was an absolute
success. We had few dropout, who have also suddenly dropped out of the
company as well. The only "complaint" was that "we are not university
students, why should we do those pointless exercises about them". But of
course it was simple to explain that Andrew would have a hard time to write
a book based on Media Gateway Controller production related code.
Anyway. In my company people cannot be forced to learn. These guys were
pretty much overloaded, and they still did show up, made the exercises and
reated the book (and because of that me) excellent. It speaks for itself.

 

[Jonathan Turkanis]

There are no changes, only "bug fixes", making the original intent clear.
Andrew Koenig has an unofficial list of changes on his web (search
comp.std.c++ for the announcement) or you can buy the PDF version of the
2003 standard at the ANSI store for 18USD.

The distinction between 'making the original intent clear' and fixing
small problems that nobody noticed can be a bit fuzzy. I thing it's
safer to say that the changes are all relatively minor, major changes
and extensions being reserved for future consideration.

Jonathan

 

[Attila Feher]

The distinction between 'making the original intent clear' and fixing
small problems that nobody noticed can be a bit fuzzy. I thing it's
safer to say that the changes are all relatively minor, major changes
and extensions being reserved for future consideration.

Yeah, it would be more politically correct to tell: "you can say" it is
only...

Actually as I have understood the changes are mainly fixes and
clarifications. And even most of the fixes are just making the text say
what it already wanted to.