X3H5 meeting review

[Fred Tydeman]

The following are notes taken by one of the people who attended the latest
X3H5 meeting.  I am just passing them on as they may be interest to some of the NCEG members.

Committee code  :ANSI X3H5
Committee title : Parallel Processing Model for High Level Programming
                  Languages
 
Dates of meeting: 9/9/92-9/11/92
Place of meeting: Palo Alto, CA
 
----------------------------------------------------------------------
 
The ANSI X3H5 committee met in Palo Alto, CA 9/9/92-9/11/92.
 
1. The model document was approved to be forwarded for public comment
   based on the letter ballot.  Votes: 9 yes, 1 no. (4 voting members
   did not respond.)
 
   (The comment attached to the no vote indicated that the model
   document was improving; but was not ready for review by people who
   had not followed the X3H5 work closely.  The lack of a companion
   binding document makes it too difficult to comment on the model
   effectively.)
 
   A significant set of editorial changes were made during the meeting.
   The model document will be distributed for review again after the
   edits are made and prior to forwarding for public comment.
 
 
2. Fortran binding.
 
   The project proposal for the X3H5 committee to do a binding of the
   model to Fortran 90 has not yet been formally approved.
 
   A list of rules were developed as guidelines to write the
   Fortran 90 binding.  They include:
 
   - Make wording consistent with a "thin binding" (ie. reference
     pertinent information in model document rather than repeat it.)
   - Do not bind to deprecated forms
   - Define new grammar and constraints in spirit of Fortran 90
   - Align section in binding with those in the model document
 
   Changes discussed in changing from Fortran 77 binding to Fortran 90:
 
   - Derived_types will be used to define locks, events, sequences
   - A SYNCHRONIZE statement will be defined with a syntax similar
     to an I/O statement.  Keywords will be used to indicate the
     "action" of lock, unlock, post, wait, etc.
   - Drop the NEW statement.  Define PARALLEL statement as new scoping
     units.  Allow specification statements after PARALLEL statement.
     (Review internal subroutines in Fortran 90 for similarities in
     documenting interface.)
 
 
3. C binding
 
   The C binding was not discussed at this meeting.
 
The next meetings are:
 
  Dec      02-04, 1992 : Tulsa, OK
  Mar      01-03, 1992 : Raleigh-Durham (?)
  Jun      14-16, 1992 : Boston
 
 
These are not IBM's views, these are the personal ones of:
Fred Tydeman, IBM, Austin, Texas (512) 838-3322; fax (512) 838-3484
AIX S/6000 Math library architect & IBM's rep to NCEG (X3J11.1)
Internet: tydemanaibmpa.awdpa.ibm.com    uucp: uunet!ibmsupt!tydeman