(j3.2006) Materials for 199

[Bill Long]

This seems superior to the cumbersome MACRO facility.  More along the 
lines of templates, which are requested with some regularity.   It seems 
more like a candidate for F201x, than for another TS.  But I don't 
really want to spend a lot of time on this at m199.  It is more 
appropriate for the Delft meeting.

Cheers,
Bill


On 9/20/12 10:20 PM, Van Snyder wrote:
> On Tue, 2012-09-18 at 20:44 -0700, Van Snyder wrote:
>
>> The simpler scheme I have in mind is to provide for definition of
>> parameterized or "abstract" subprograms, which cannot be invoked
>> directly....
>
> I don't know whether this meets Malcolm's "really**7 simple" criterion;
> maybe it's only "really**4 simple."  I've put together a description of
> abstract subprograms and how they're used.  It's in the format of a TR,
> complete with the repetitive top-down development.
>
> Counting cover page, frontispiece, ISO boilerplate, introduction,
> foreword, it's 14 pages.
>
> There are almost four pages of narrative description, and just over
> three pages of draft edits.
>
> My estimate that it would add fewer than five pages to the standard
> appears to have been an over estimate.  It looks like fewer than two,
> because it introduces only a few new ideas, and exploits mechanisms and
> concepts we already have in place.  If the examples in the paper are put
> into Annex C in the standard, that would add one more page.
>
> During F08 development, at least at the beginning, we had a 010 document
> that listed paper numbers for specs, syntax and edits.  This quickly
> became irrelevant, as we dribbled in pieces of features a screw here, a
> rivet there, some bubble gum, some duct tape....  It might make for more
> coherent development if we kept a coherent "project" paper for each
> feature until we are quite convinced it's complete, as we would for a
> real TR or TS, but without the ISO boilerplate.  The 010 document would
> list the current version.  Each version should list the paper numbers of
> previous versions in a "history" section.  This precludes frequent
> revision of 007, unless the champions for features agree to go through
> accumulated edits and update them when it comes time to put the feature
> into 007.
>
> If projects interact we could insert UTI's in all of those that
> interdepend upon one another, so as not to forget integration issues.
>
>
>
>
> _______________________________________________
> J3 mailing list
> J3 at mailman.j3-fortran.org
> http://mailman.j3-fortran.org/mailman/listinfo/j3
>

-- 
Bill Long                                           longb at cray.com
Fortran Technical Support    &                 voice: 651-605-9024
Bioinformatics Software Development            fax:   651-605-9142
Cray Inc./Cray Plaza, Suite 210/380 Jackson St./St. Paul, MN 55101