(j3.2006) [Re: Spreadsheet from meeting 167]

[David Muxworthy]

> despite BSI schizophrenia of presenting a tutorial on coarrays at
> Trollhattan in 1998, between J3 meetings 145 and 146, asking WG5 for
> them officially at 166 (UK-001), and then asking at 186 to remove them
> from the work plan, after the project was completed!

In defence of BSI, in 2005 we (BSI) voted for the principle of
coarrays being incorporated in F08.  Unfortunately it seems that most
of us hadn't quite realized at that stage what a huge impact it would
have on the language.  Also we had been led to believe that, being a
more elegant solution, coarrays would soon prove to be more popular
with users than MPI and OpenMP.  Already by 2006 I personally thought
WG5 had made a mistake and that coarrays should be in an optional part
of the standard, but of course that idea got nowhere with WG5.  Even
people who told me privately that they agreed with me still voted to
preserve the status quo.  By 2008 more of the BSI group had come round
to the view that too much effort was being expended for a small
minority of users.

> Meanwhile, why did we piss away a goodly chunk of the last five meetings
> working on Dan's half-baked favor for a sister working group, that WG5
> promised to do even though it has almost nothing to do with Fortran,
> developing NON-NORMATIVE material that ought to be in textbooks, not
> standards?  At Garching, it was agreed that Dan, John, David and Makki
> would finish it.  The only reason I did not oppose it at Garching is
> that it was clearly advertised as a WG5 project.  Why is J3 doing all
> the work?

As I understand it the decision at Garching was based, without any
great enthusiasm, on the belief that collaborating with WG23 was
better than ignoring them.  I am appalled to hear that J3 has spent
significant time on the Fortran annex.  The annex is essentially a
political document which few will ever use for technical work.  The
drafting could certainly be done by the WG5 subgroup by email.

David