[J3] [SC22WG5.6409] Kickstarting proposals for F202Y features

[Van Snyder]

On Tue, 2022-07-05 at 14:00 +0000, Jeff Hammond wrote:
> If you think we should limit Fortran 2026 to what Ada was willing to
> do in 1983, I don’t know what to tell you.  Fortran isn’t Ada and it
> has been almost 40 years since that decision.  Many aspects of
> computing have changed and we do things today that were almost
> unimaginable in 1983.

Like Fortran, Ada is continuously under development. The 1983 Ada
standard has been revised many times. The most recently published
revision was in 2012, when full support for "programming by contract"
and "protected variables" (aka coroutines with synchronization) were
provided. Another revision is nearing publication. A draft of the 202x
standard is available at http://www.ada-auth.org/standards/ada2x.html.

Ada wasn't designed for linear algebra. It was designed for embedded
real-time applications. So their original approach was for unstructured
parallelism, not coarrays. Now that Fortran is also finally thinking
about less structured parallelism, not just linear algebra and PDEs, it
would be useful to learn from the Ada community's very long experience.
Read "Concurrent and Real-Time programming in Ada" instead of guessing
that all that's there was what was provided forty years ago. They have
actually learned something about what we in Fortran are only now
starting to imagine.

If anybody is interested in details about why async blocks were
rejected by the Ada community, in favor or protected variables (aka
coroutines with synchronization), I can put you into contact with
members of WG9.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20220705/24bc5ff6/attachment.htm>