🚀 go-pugleaf

(This is the third part of this message.)

        Q: Why doesn't C9X include feature X that exists in C++, ``you
        can't write good applications unless you include feature X''?
        A: Feature X, whatever that is (e.g., inline functions,
        operator overloading, function overloading, templates, etc.),
        might be considered if there was much support AND it was
        well-defined AND it is light weight.  Having reviewed several
        drafts of the C++ Standard (it has failed CD registration ballot
        recently), it is easy to see why many things aren't defined well
        in C++.  See below.

        Q: If the C++ Standard is coming out soon, why do we need these
        features in C9X?
        A: C has a different market than C++.  C still caters to
developers
        of: systems programming, high performance systems, embedded
        systems, numerical applications (see the latest Technical
        Report that added numeric features).  Also, the C++ Standard
        approaches 1000 pages *without* a rationale section.  Based upon
        the size and complexity of the C++ standard, it will take them
        at least 4-5 years to solve these problems.  It isn't because
        the C++ committee lacks bright people (there are many bright
        people in WG21), it is that this is the nature of language
        standards.  Each little feature that gets fixed produces several
        second and third order effects.  It takes many years for features
        to ``bake''.  The main problems with the C++ Standard that drag
        down the process are:

                (1) Too many new, untested features added to the language.
                (2) No rationale -- Stroustrup's book on the evolution
                of C++ is incomplete and inaccurate (my source of info
                is based upon others with first hand experience in the
                committee).
                (3) Too much work to do.  IMO, C++ should have
standardized
                the language first (with minimal library support), then
                produced a technical report or an amendment with the
                libraries later on.  In WG14, we had much success with
                this approach in Amendment 1 (included the multibyte
                support extension library) and the Technical Report
                (included many features useful for numeric programming).

        In summary, C might have a standard that includes class-like
        features sooner than C++.

        Q: Did class-like features just get introduced in C9X?
        A: No.  This has been an ongoing discussion since the 1993-12
        meeting with the first paper on adding classes to C.  Much of
        the discussion over the past two years has been what subset
        are we interested in.  We know many C++ programmers only use
        a small portion of the language to write useful programs.  See
        below.

        Q: What is the relationship to Java?
        A: Java is based upon C, not C++.  Java has many of the features
        proposed above.  Although C9X is not standardizing Java (Java
        doesn't allow pointers, for example), I would think that anyone
        with interest in Java would want to participate in WG14 since
        they cover about the same subset in class-like features.


        Q: How do I participate?
        A: Contact your national body (i.e., the standards organization
        in your country).  In the US, it is X3 (phone +1 202 737 8888).
        You might also contact Rex Jaeschke, X3J11 chair, via E-mail
        at "rex@aussie.com".  IMPORTANT: Before you consider participating
        in the process, consider your capabilities as a company.  Meetings
        are held for 5 days every 4 months (February, June, October).
        Usually, one meeting out of three (usually the June meeting) is
        held outside North America since the meetings are co-located
        (not joint) X3J11 and WG14 meetings.  You should have the
financial
        resources to attend meetings.  You need to attend 2 out of 3
        meetings so you can skip the non-North America meetings if your
        finances don't support this.  The C9X project is expected to
        complete by 1999-12.  You need think about participating NOW
        since the last 1-2 years of the project are public review and
        balloting.  If you can't attend the meetings, you can still get
        the mailings -- see X3 about observer status.  Finally, as
        individuals you are always able to supply comments to your
        national body.  You should consider this if you don't have the
        financial or workload capability.  X3J11 and WG14 have been
        pretty good about responding to the public comments and
        suggestions.  Typically, official responses to suggestions are
        about 6-18 months and 1-2.5 years for defect reports.  Usually,
        you'll have some initial feedback from your national body.

        Q: When are the next meetings?
        A: The following is the meeting schedule for the next three years:

                1996-02-05/1996-02-09  Irvine, CA         Unisys
                1996-06-24/1996-06-28  Amsterdam          NNI
                1996-10-21/1996-10-25  Toronto            IBM
                1997-02-03/1997-02-07  Kona, HI           Plumhall
                1997-06-23/1997-06-27  Syndey, Austrailia Plauger
                1997-10-20/1997-10-24  Silicon Valley     Sun
                1998-02-??/1998-02-??  ???                ???
                1998-06-??/1998-06-??  ???                ???
                1998-10-??/1998-10-??  New York City      Farance

        Q: Are classes now a part of C9X?
        A: No.  They are being considered for C9X.

-FF
-------------------------------------------------------------------
(``I only use AOL for reading netnews.'')
Frank Farance, Farance Inc.
E-mail: frank@farance.com, Telephone: +1 212 486 4700
ISO JTC1/SC22/WG14 & ANSI X3J11 (C Programming Language) Project Editor