[comp.lang.c++] Lippman's C++ Primer vs. Ellis' ARM SUMMARY

rclark@concour.cs.concordia.ca (CLARK richard) (06/21/91)

    Hello all,

    I wasn't necessarily planning on posting the results of my survey,
    but I received so many fine answers that I felt I should share them.

    I've received numerous replies about my query which basically asked
    if the Annotated C++ Reference Manual was worth purchasing if I already
    owned Lippman's C++ Primer (which I considered to be quite complete, but
    again, it's the only C++ book I've read).  I must thank this newsgroup for
    pointing out that book to me too.

    The reasons for asking were :

    1) I felt I needed something a little more than the book by Lippman.
       It was sometimes hard to find exactly what I wanted, and I was
       running into problems coding, saying "Let me get this straight
       ..."  I wanted to know if the ARM provided more information, or
       put it across differently.

    2) I had seen numerous references to the ARM, but my computer
       bookstore didn't have it in stock - I would have to order it
       without seeing it.

    [As an aside, I've ordered "the ARM"]

    I would like to thank each of the respondants for the time and
    trouble it took to send me some information.  They are (in no
    particular order) :

     Daniel A. Burkhard     Russ Williams      Frank Griswold
     Ron Schweikert         Kim Coleman        Gary Duzan
     J. W. Hargrove         Stanley Friesen    Steven Parkes
     Keith Walker           Mike Oltz          Mike Lijewski
     Robert Jacobs          Amitabh Agrawal    Philip Machanick
     Stephen Clamage        Randy Hudson       Paul Lyon
     Gary L. Randolph       Michael Cox        Jamshid Afshar
     Joe Pillera            Bob Brazile        Martin Hitz
     marty hoffmann         Matt Mahoney

    SUMMARY :

    A large majority considered the "Annotated C++ Reference Manual" by
    Ellis and Stroustrup to be _the_ definitive source for information
    on C++, and that it contained information in a format much easier to
    find than in the book by Lippman because it was in reference manual
    format (which is natural, Lippman's book is a tutorial).  The ARM
    goes into more intricate detail explaining the "why's", and
    sometimes the "how's" of an implementation.  The ARM is _not_ the
    book to learn C++.

    The book by Lippman is an excellent introduction to C++, but it
    can't be everything to everyone.  For many C++ programs, the book by
    Lippman is sufficient, but for more advanced programming the
    Reference Manual is required.  It goes into detail that a tutorial
    can't.

    No other book was mentioned as an alternative to the ARM.  I thought
    there might be more than one reference manual, for example for C we
    have books by Kernighan & Ritchie, and Harbison and Steele.

    Following are some of the points that were made.  I've edited the
    responses significantly and tried to avoid duplication, so if
    you see 3 negative things and 3 positive things, that doesn't mean
    opinion was split 50/50.  Opinion was definitely that owning
    the book by Lippman should not preclude purchasing "the ARM".

    I'll not identify the authors individually (you know who you are).

    "It [The ARM] is not at all tutorial, but a significant number of
annotations are provided to explain what may not be obvious from the
strict language description."

    "I used Lippman to learn C++ -- I now have the ARM and really never
look at Lippman.  Lippman is good for learning but does not have enough
detail to be a good reference for serious programming."

    "Lippman has lots of examples where you can easily see which
headers to include. Since the libraries aren't (it would seem) part of
the standard, this is an area not covered at all in E&S"

    "Ellis & Stroustrup is as close as one will come to an "official"
definition of the language."

    "On the other hand, the ARM alone is of little help, as a reference
manual it lacks a 'thread`, it consists to 40% of cross references."

    "You really have to know C++ before you can use it [The ARM]."

     "It [The ARM] isn't *really* necessary other than for a final
reference."

     "[The ARM] is sort of like "Everything You Ever Wanted to Know
About C++, But were Afraid to Ask".

    "I have not found Lippman's book to be useful as a reference, and I
have not found E&S to be useful as a tutorial.  Each is excellent for
its intended purpose, however."

    "[The ARM] is mainly good for answering the question: "Is this
legal?"

    "both Lippman's C++ Primer and Stroustrup's C++ Programming Language
(his OTHER book) are coming out in new editions next month.  Sigh.  But
the ARM is not being revised real soon, as it already covers up to C++
3.0."

    "However, I would recommend ARM only if you are not using g++. G++
does not strictly follow the language definition."

    "The book *Programming in C++* by Stephen C. Dewhurst and Kathy T.
Stark is a good sequel to the Primer."

    "If for no other reason you'll need to when someone references it
and a page number in comp.lang.c++."

    ""Depends on if you program right up to the limits of the language.
(and of course, you will :-)"

--

rick clark (rclark@concour.cs.concordia.ca) | Membre du "Bloke Quebecois"
Concordia University, Montreal.             |