(j3.2006) (SC22WG5.5983) [Fwd: ISO/IEC JTC 1/SC 22/WG 5 New documents available on ISOTC eCommittee]
Van Snyder
Van.Snyder
Mon Oct 9 14:51:33 EDT 2017
On Sun, 2017-10-08 at 17:05 +0200, Steve Lionel wrote:
This is the current summary of the Fortran 2020 Feature Survey that WG5
has been running since July. It will be updated again in January to
reflect any new responses. The purpose of the survey is to collect
comments and suggestions from the Fortran user community on features
they would like to see in the next Fortran standard. No obligation to
implement any of the suggestions is implied.
N Number
Title
(Description)
Exp. Action
Due Date
Version Date
2142
Fortran 2020
Feature Survey
Results
2017-10
INFO
None
2017-10-08
A few comments on the comments:
There's a lot of hidden duplication.
There are many requests for things we've been asked to do many times
before, such as short-circuit logical operators, default initial values
for absent arguments, automatic initialization without SAVE, first-class
enumeration types....
Some interesting things are well hidden. For example, one responder
requested iterators under the rubric of generic programming. Iterators
are to coroutines as functions are to subroutines. We shouldn't do one
without the other. If we had added coroutines when we added unlimited
polymorphic, and had allowed to specify a coroutine in an I/O statement
to handle list elements, we might not have done UDDTIO. A little bit
more forward thinking would be useful.
Another interesting one was extensible modules. Submodules are
extensible. They were consciously based on Ada private child units.
Ada has had public child units (read "extensible modules") since 1995.
This was proposed during discussion of submodules, but rejected.
Units of measure were requested several times, but weren't on the
"what's your rank for this" list. At least one responder suggested
publishing trial balloons as TS's. Proposals to publish a non-binding
TS on units, i.e., a trial balloon, have been vociferously resisted.
More information about the J3
mailing list