(j3.2006) (SC22WG5.4192) WG5 Las Vegas Feb-2010/J3 m191 minutes 10-144

Van Snyder Van.Snyder
Tue Feb 23 20:06:30 EST 2010


On Tue, 2010-02-23 at 13:55 -0800, Bill Long wrote:

> As Malcolm has already noted, past predictions of a "maintenance only"
> update of Fortran have always turned out to be substance-free...

Malcolm made that remark in the context of Fortran 95.  I think that's
the only revision that was intended to be "maintenance only."  After
Fortran 90, however, there were two TR's and three corrigenda, so the
development agenda for Fortran 95 was at least to incorporate them.  I
don't know whether the initial plan was to do only that, or whether it
included the five deletions and two semantic changes listed in 97-007r2
subclause 1.5.1, or whether it included FORALL, PURE and initialization
as outlined in the 97-007r2 introduction and whatever other extensions
were too minor to mention there.

> ...  If the goal is to get coders under the age of 50 (arbitrary
> dividing line) interested in Fortran, then I think that enabling OOP
> and making Fortran a natively parallel language are both winners. At
> least that's been my experience so far.  Somehow, I don't think that a
> physical units facility has comparable sex appeal.

The only reason OOP wasn't on my 2003 wishlist (97-114) was that it was
already on the 2003 workplan when I joined J3 in February 1997.  The
only reason coarrays weren't on my 2008 wishlist is that I knew from the
1998 WG5 meeting that they would be on the British work plan request.  I
have had different parallelism features for multiprocessor shared-memory
systems on my wishlist since 97-114.

Although physical units, coroutines, a more unified and complete type
system, uniform syntax, ... don't have the sex appeal of OOP and
coarrays, they are also far less work, perhaps in toto, than either one.

> I'm sure there will be a wish list for the next revision that is too
> long, and we (WG5) will have to make choices, as always.

That is certainly true, eminently reasonable, and to be expected.

My sponsors have made it clear, however, that they have an agenda that
goes beyond finishing what was pushed off the original 2008 work plan in
order to meet the schedule.  If the WG5 edict becomes "no new features
other than currently existing TR's until 2018," I can guarantee that JPL
will stop paying my INCITS dues.  My immediate management controls my
(domestic) travel funds, while somebody several levels up controls my
INCITS membership funds (and maybe NASA HQ is involved).  I have to
write a four-page memo every year to justify the latter.  If I can't
vote, my immediate management might well decide not to send me to
meetings.  





More information about the J3 mailing list