Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming.threads > #393

Re: The threading specs in the standard: a new catastrophe

Cross-posted to 3 groups.

Show key headers only | View raw

Mr Kuyper

If simply supporting pthreads is not so simple as I described (and
I fear you are right) why don't we leave that part out?

Why this rush to standardize another, competing standard in a highly
complex subject matter where most committee members aren't experts?

Can't we postpone this whole subject for a later edition? Why must
the committee do something in this direction?

Looking at the situation now and the errors that are discovered when we
look deeper into this, makes obvious that this whole business is a
VBI: a Very Bad Idea (tm).

Without trying to denigrate anyone it is obvious that most committee
members aren't experts in this thread stuff, as many people, there is
*nothing wrong* with that. They CAN'T KNOW EVERYTHING.

Both windows threads and POSIX pthreads are the work of specialists
of the field, and they are standards pushed by a company that has
enormous resources or by a community of users with thousands of
developers that have tested and explored the pthread design since YEARS
and YEARS.

To believe that the C standards committee with so few resources at
hand and with no significant user base at the start will succeed
is beyond what I can sanely conceive. This will take a lot of
time and an enormous effort from the committee members for dubious
results AT BEST.

Isn't a revamping of the standard library (that hasn't been really
updated since the seventies) more important for the C language?

Personally I just do not understand why this comes up. True, C++
has started a similar approach but they have a larger user community
and probably more industry support. And even in their case, any
new standard that would present the same glaring defficiencies
as the proposal of the draft we are discussing would be largely
ignored by a community where care of details is VERY important.

Designing multi-threaded programs is notoriously difficult and
people will be VERY reluctant to change the base support they
are working with since years. Specially if it is fundamentally
flawed. If you do not give the implementors the possibilty of
using ALL (or the most crucial) features of the underlying
OS this C API will never take off.

Please consider this viewpoint

Yours sincerely

jacob
-----------------------------------------------------
P.S. Concerning your question about static initialization,
ANY static initialization can be dynamically simulated by
code that runs before main() starts. Specifically

int myInt = 42;

can be simulated by

_cStartup() // this is the entry point
// ...
myInt = 42;
// ...
return main(argc,argv);

For the program, it is AS IF that integer was statically initialized.

Back to comp.programming.threads | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Re: The threading specs in the standard: a new catastrophe jameskuyper <jameskuyper@gmail.com> - 2011-07-07 15:25 -0700
  Re: The threading specs in the standard: a new catastrophe sfuerst <svfuerst@gmail.com> - 2011-07-07 16:12 -0700
    Re: The threading specs in the standard: a new catastrophe James Kuyper <jameskuyper@verizon.net> - 2011-07-07 22:41 -0400
    Re: The threading specs in the standard: a new catastrophe "Chris M. Thomasson" <cristom@charter.net> - 2011-07-08 12:47 -0700
      Re: The threading specs in the standard: a new catastrophe sfuerst <svfuerst@gmail.com> - 2011-07-08 13:11 -0700
        Re: The threading specs in the standard: a new catastrophe "Chris M. Thomasson" <cristom@charter.net> - 2011-07-08 16:56 -0700
        Re: The threading specs in the standard: a new catastrophe Marcin Grzegorczyk <mgrzegor@poczta.onet.pl> - 2011-07-09 23:55 +0200
          Re: The threading specs in the standard: a new catastrophe sfuerst <svfuerst@gmail.com> - 2011-07-09 18:40 -0700
  Re: The threading specs in the standard: a new catastrophe jacob navia <jacob@spamsink.net> - 2011-07-08 01:16 +0200
    some sane words, for sure "Greg A. Woods" <woods@robohack.org> - 2011-07-07 18:33 -0700
    Re: The threading specs in the standard: a new catastrophe jameskuyper <jameskuyper@gmail.com> - 2011-07-08 08:34 -0700
      Re: The threading specs in the standard: a new catastrophe sfuerst <svfuerst@gmail.com> - 2011-07-08 10:47 -0700
        Re: The threading specs in the standard: a new catastrophe James Kuyper <jameskuyper@verizon.net> - 2011-07-09 06:57 -0400
          Re: The threading specs in the standard: a new catastrophe Alexander Terekhov <terekhov@web.de> - 2011-07-09 13:59 +0200
            Re: The threading specs in the standard: a new catastrophe jacob navia <jacob@spamsink.net> - 2011-07-09 14:45 +0200
            Re: The threading specs in the standard: a new catastrophe "Chris M. Thomasson" <cristom@charter.net> - 2011-07-10 15:47 -0700
              Re: The threading specs in the standard: a new catastrophe Alexander Terekhov <terekhov@web.de> - 2011-07-11 13:38 +0200
            Re: The threading specs in the standard: a new catastrophe Phil Carmody <thefatphil_demunged@yahoo.co.uk> - 2011-07-11 16:12 +0300

csiph-web