C++ Logo


Advanced search

Re: [wg14/wg21 liaison] Designated initializers in C++ and C

From: Bjarne Stroustrup <bjarne_at_[hidden]>
Date: Thu, 13 Aug 2020 15:19:21 +0200
On 8/13/2020 11:08 AM, Uecker, Martin wrote:
> To me a merger seems difficult. I fear that C would loose its
> identity and essentially cease to exist.
> C is is maintained by a separate working group, has
> different use cases, and serves a different group of users.
> I can not see that this can work if it also is required
> to be a subset of C++.

A merger would obviously be difficult. There are differences in
overlapping - but incompatible - feature sets and in the text of the
standards. Two decades ago, I thought a merger would be technically
feasible; now I am not so sure.

  * B. Stroustrup: C and C++: Case Studies in Compatibility
    The C/C++ Users Journal. September 2002. Pdf version
  * B. Stroustrup: C and C++: A Case for Compatibility
    <http://www.ddj.com/dept/cpp/184401555>. The C/C++ Users Journal.
    August 2002. Pdf version <https://www.stroustrup.com/compat_short.pdf>.
  * B. Stroustrup: C and C++: Siblings
    The C/C++ Users Journal. July 2002. Pdf version

That was 20 years ago; much have changed since.

People talk about C's "separate identity" and (as above) C "has
different use cases, and serves a different group of users" from C++. I
have never seen C's separate identity clearly articulated, nor have I
seen articulation of fundamental reasons for there being separate use
cases. Such articulation might simplify the discussions about which
incompatibilities are deliberate and essential and which are accidental.

For attempted characterizations of what C++ is and should ideally be, see

  * Bjarne Stroustrup: Thriving in a crowded and changing world: C++
    2006-2020 <https://dl.acm.org/doi/abs/10.1145/3386320>. ACM/SIGPLAN
    History of Programming Languages conference, HOPL-IV. London. June 2020.

    H. Hinnant, R. Orr, B. Stroustrup, D. Vandevoorde, M. Wong:

On the other hand, I consider the value of compatibility obvious. Even
in this thread, people have been surprised about differences in the

Sometimes, my comments about compatibility are taken as attacks on C or
displays of ignorance. They are not. I built C++ on C and few C
programmers today have written a C program that doesn't critically
depend on some of my work. I was a friend of DMR and still see BWK
fairly frequently. The compiler used for the code in K&R2 was written by
me - and all the code was in the C/C++ common subset. However, I do not
think that either C or C++ are perfect, so I comment on flaws; such
comments are not attacks on C or C++ (though of course they could
reflect misunderstandings), but the first steps to potential improvements.

Received on 2020-08-13 08:22:49