TITLE: Notes on the ARM and D&E

RESPONSE: bs@alice.att.com (Bjarne Stroustrup), 1 Apr 94


Let me try explain how the ARM, ``The C++ Programming Language (2nd edition)'',
and my new ``Design and Evolution of C++'' books relate to each other.

The ARM

	Margaret A. Ellis and Bjarne Stroustrup:
	The Annotated \*C Reference Manual.
	Addison-Wesley, Reading, MA. 1990.
	ISBN 0-201-51459-1. 478 pages. Hardcover.


	is the C++ reference manual annotated with information supposedly
	of interest to language lawyers and implementers, plus a chapter on
	ANSI/ISO resolutions.

	Eventually, the ARM and every other C++ reference manual will be
	outdated by the work of the ANSI/ISO committee. For most users,
	I don't think this has happened yet. If you write C++ compilers,
	C++ test suites, and the like you need to have the committee's
	working paper (and preferably be represented on the committee);
	if not, the ARM provides a sufficient - and importantly, stable -
	reference.
	

My 2nd edition

	Bjarne Stroustrup:
	The \*C Programming Language (2nd edition).
	Addison-Wesley, Reading, MA. 1991.
	ISBN 0-201-53992-6. 696 pages. Softcover.


	is a complete C++ tutorial for experienced programmers,
	plus the C++ reference manual,
	plus three chapters on design issues,
	plus a chapter on ANSI/ISO resolutions.

	You can learn C++ programming from the 2nd edition, but not from
	the ARM (not easily, at least). You can find really obscure details
	and techniques for implementing C++ in the ARM, but not in my 2nd
	edition. My 2nd edition has a chapter on iostreams (from a user's
	point of view). Neither book describes any of the larger C++
	libraries or any implementation specific details.

	The chapter on ANSI/ISO resolutions is common to the ARM and
	my 2nd edition (modulo differences between printings).
	It is available via ftp

		ftp ftp.std.com
		cd AW
		cd stroustrup2e
		get ...

	I also place errata there.

	The ANSI/ISO chapter is an attempt to address the problem of how to
	keep a book completely up-to-date without messing about with its
	content to the point where it becomes useless as a common point of
	reference.

	The approach is feasible because the changes fit into the overall
	pattern of the language.

D&E

	Bjarne Stroustrup
	The Design and Evolution of C++
	Addison Wesley, Reading, MA. 1994.
	ISBN 0-201-54330-3. 472 pages. Softcover.


	is a discussion of why C++ is the way it is. It describes the
	design of C++. The emphasis is on the overall design goals,
	practical constraints, and people that shaped C++. It is not a
	book trying to document every little detail of the language or
	its use. The point of view in the discussion of C++ tend to be
	that of a user (programmer or designer) rather that of a language
	lawyer.

	Most books you find in a bookseller's ``programming languages''
	section fall into one of two categories: The manuals (``this is
	what XXX is'') and the tutorials (``this is the way you do YYY'').
	D&E does not fall into either. It relies of my 2nd edition (or some
	other quality tutorial) for teaching C++ programming, and it is
	certainly not a manual (use the ARM or the reference manual in my
	2nd edition).

	D&E should give you a much better appreciation of the way the
	features of C++ fit together. It is my hope (and experience) that
	understanding of such matters help people write better programs,
	but the book's primary aim is not on specific programming techniques.

	Naturally, D&E is completely up-to-date in its description of C++.
	The features described in the ANSI/ISO chapters of the ARM and my
	2nd edition are described in some detail in their proper contexts
	in D&E.

The overlap in the content of the three books is quite limited. The ARM
and my 2nd edition share the reference manual and the ANSI/ISO chapter.
The D&E and the ARM annotations overlap slightly (on the order of 10%
I would guess) and D&E and my 2nd edition share a few examples. However,
even in those cases the point of view tend to differ. I'm pretty sure
that the overlap in contents between any two quality tutorials is far
greater.