From martin@ocfmail.ocf.llnl.gov Fri Jul 17 02:17:28 1992
Received: from [134.9.48.4] by dkuug.dk via EUnet with SMTP (5.64+/8+bit/IDA-1.2.8)
	id AA11220; Fri, 17 Jul 92 02:17:28 +0200
Received: by ocfmail.ocf.llnl.gov (4.1/SMI-4.0)
	id AA28299; Thu, 16 Jul 92 17:17:41 PDT
Date: Thu, 16 Jul 92 17:17:41 PDT
From: martin@ocfmail.ocf.llnl.gov (Jeanne Martin)
Message-Id: <9207170017.AA28299@ocfmail.ocf.llnl.gov>
To: SC22WG5@dkuug.dk
Subject: Summary on N760 Comments (N792)
X-Charset: ASCII
X-Char-Esc: 29


To:       WG5
From:	  Jeanne Martin
Subject:  Summary of N760 Comments

In order to speed the development of a strategic plan, I have done three things:
First I summarized the comments on N760.  The summary is N792.  I apologize
if I have misunderstood or misrepresented any comments.  Second, in order to
focus the discussion in Victoria, I developed, from the summary, several straw
votes.  The votes are in N793. Third, because it takes a long time to revise
the document and make it consistent throughout, I made guesses at the outcome
of the votes and created a proposed replacement document for N760.  It is
N794.  When creating N794, I tried to take into account the suggested editorial
changes as well; however, this document is the work of only one person and has
no more official status than any other proposal that is prepared for
consideration at a meeting.  I invite anyone else to create such a document for
equal consideration.  If you have comments or suggested changes, I'm sure
everyone would like to consider them ahead of time.  However, I will not be
able to revise either N760 or N794 again before the meeting.  Time is growing
short.


                                                 ISO/IEC JTC1/SC22/WG5 - N792

               	       SUMMARY of N760 COMMENTS

UK comments:
   - Suggests removing references to project editors.  [Since project editors
     are an SC22 requirement, this will not be done.]
   - Suggests changing schedule to allow more time for requirements and less
     for development.  Suggests allowing the processes to overlap.
   - Suggests annual corrigenda rather than one in 1995.
   - Suggests that X3J3 be invited to fulfil the role of primary development
     body.

US comments:
   - Requires that bodies coordinating their work closely with the primary
     development body be able to standardize syntax and semantics for future
     Fortran development.
   - Requires we make no firm promises about not invalidating collateral
     standards in following revisions.
   - Suggests that a revision limited to corrections and clarifications be
     produced in 1995 rather than a corrigendum.

Hoffert comments:
   - Requires that the WG5 mechanism for interim decision making not be limited
     to a preselected management committee, but that instead there be more
     frequent meetings or a more open mechanism.
   - Endorses the "train model" where there is a specified revision cycle
     (fairly short) in order to be more responsive to changing technology
     and to allow technical work and procedural processing to proceed in
     parallel.

Lahey comments:
   - Requires that X3J3 work directly under the direction of WG5 and be more
     responsive to changing technology by shortening the revision cycle.
     [I believe his aims could be accomplished if X3J3 produced an early
     revision (not strictly limited to corrections) as an "I project".]

Martin comments:
   - Suggests a short-term revision (not strictly limited to corrections)
     be produced by X3J3 as an "I project".  [Note that this means that X3J3
     produces an international standard and there is only one set of rules.]

Pollicini comments:
   - Suggests there be no amendments; candidate features should be integrated
     into revisions instead.
   - Suggests that collateral standards be prohibited from using features
     that rely on storage association.

Schonauer comments:
   - Suggests we be more responsive to changing technology by producing
     amendments between major revisions, in particular to accommodate
     parallel processing primitives and compiler directives.  [I assume
     he would concur with a shorter revision cycle.]

Schonfelder comments:
   - Suggests that it be clear that amendments will be integrated into the
     first following revision.
   - Opposes syntactic and semantic extensions being standardized except
     by WG5 and X3J3

Tait comments:
   - Requires that all references to collateral standards be removed.
   - Requires that the development body be the manager of development - not WG5.
   - Requires that the plan specify how the work of other organizations
     (not necessarily standards bodies) be managed (possibly as task groups)
     and their work incorporated into Fortran revisions. 
   - Requires an effective WG5 management committee for interim decisions.
   - Requires a mechanism for incremental development.
   - Endorses the "choo-choo train model" with a four-year cycle (1996 and 2000)