psmith@mozart.uucp (Presley Smith) (10/16/88)
In the past few weeks, there have been several reports posted on the
network about the WG5 meeting in Paris. Various people have commented
on that meeting. The detailed resolutions passed at that meeting have
been published in comp.lang.fortran by Carl Burch and Dietrich Wiergandt.
Before I discuss what happened in the last week, I want to clarify the
relationship between the various committees. I hope the following diagram
will help you understand the relationship between the various groups:
U.S. Domestic Organization International Organization
------------ ------------
| ANSI | | ISO |
------------ ------------
^ ^
| |
------------ ------------
| X3 |<------- | SC22 |
------------ | ------------
^ --------- ^
| | SPARC | |
------------ --------- -----------
| X3J3 |-------^ | WG5 |
------------ -----------
(SPARC is the group that directly oversees the work of the various committees,
reviews the committee's directions and actions, and reports their findings to
the X3 Committee. )
X3J3 produced the initial draft for FORTRAN 8x. This draft was in public
review from October, 1988 to February, 1989. During that time WG5/SC22 also
held a public review on this draft. The X3J3 public review was largely
negative whereas the SC22 review was more positive. There are reasons for
this difference.
At the Paris meeting of WG5, the resolutions detailed in the articles mentioned
earlier were passed, and the effort to produce a new draft passed back to X3J3.
The WG5 resolutions requested that X3J3 produce another version of the FORTRAN
8x standard and have this new version back to WG5 in the December, 1988
timeframe. The WG5 resolutions also imply that WG5 will take this draft
and produce an international standard for FORTRAN whether the U.S. approves
this standard or not. This raises the possibility that there might be two
different standards for FORTRAN: an ISO standard and an ANSI standard.
Such a split would be a serious situation for the FORTRAN world. ANSI rules
call for any major change in direction by any group to be reported
immediately to the X3 Committee. X3 is responsible for the management of
the X3J3 Committee. X3 is also the link between X3J3 and either the ISO
or ANSI groups. All correspondence between ISO and ANSI must pass through X3,
and X3 is responsible for monitoring this correspondence.
The resolutions passed at the WG5 meeting, and the issue of what procedures
X3J3 would follow was debated in the X3 Committee meeting in Washington, D.C.
last week and the following direction was issued to the X3J3 Committee:
------------ Exact Text of the X3 Direction to X3J3 ----------------------
X3 Direction to X3J3
The following motion was approved by a vote of 19-0-1
X3 appreciates the diligent work of X3J3, however, X3 is concerned with the
pace of development of a new Fortran standard.
X3 directs X3J3 that:
1) The U.S. support a single Fortran standard
(domestic/international, i.e., one document),
2) All public review comments (and the comments contained in
SC22/WG5, Resolution 2) should be considered in the normal
X3 procedures before transmitting any revised draft
standard to SC22/WG5. X3J3 will notify WG5 of its proposed
work plans and schedule.
3) The Chair of X3J3 submit a report of the plan and progress
to SPARC after the X3J3 Boston meeting and each subsequent meeting.
--------------- End of X3 Text ---------------------------------------
What follows is my assessment of the situation and what this X3 direction
means to both X3J3 and WG5:
1. The U.S. does NOT want two standards. Because X3J3 is the group
responsible for producing the draft standard, it is X3J3's
responsibility to solve this problem.
2. X3J3 is governed by a set of rules defined in the SD-2 document,
"Organization and Procedures", and X3J3 must follow those
procedures before transmitting any new document to SC22/WG5.
Currently, X3J3 is in Milestone 14 of the SD-2 which states:
"When there are X3 negative votes, adverse public comments or
comments accompanying affirmative votes, the voting tally and all
comments are sent to the originating TC (technical committee, X3J3
in this case). The TC considers each comment and prepares a response
to each in accordance with the provisions of the ANSI BSR."
Procedure for answering public comment is:
"The comment should be discussed at the next TC meeting, and if
not definitively responded to at once, an interim acknowledgment
should be sent along with an estimated date of action." (Such
letters have been sent to those who sent in comments.) "When
a final response is issued you must inform the commentors of their
need to notify the Secretariat of their satisfaction or
dissatisfication with the committee's response. The commentor is
required to send the Secretariat a written statement indicating
acceptance or rejection of the TC response within fifteen working
days. The commentor must be made aware that failure to respond
within fifteen working days that the comment stands will indicate
to the Secretariat that the comment is to be withdrawn."
Because X3J3 has not yet written the 400 letters to the public review
commentors, those commentors have not had their 15 working days
to respond to the committee's position, etc. procedurally it
would be impossible to produce another version of the proposed
standard on the schedule contained in the WG5 resolutions.
This effort and certain other actions must be completed prior to
X3J3 ballot on any new proposed standard and a re-transmittal to
WG5 per the X3 direction.
The X3 direction to X3J3 is to follow the SD-2 rules and to notify
WG5 of the program of work and schedule to complete this effort.
3. It is clear that X3 intends to monitor this process more closely.
The process of having the Chair of X3J3 report progress to SPARC
after each meeting will further focus attention on the situation
and should help X3J3 come to a more timely completion of the
process of producing a new FORTRAN standard.
There is an SC22 AG meeting in Tokyo the week of 10/17. The Paris
resolutions of the WG5 group and the X3 action will be discussed at
that meeting. I would expect a reply to the X3 action and additional
direction to WG5 to come from that meeting.
The next meeting of X3J3 will be the 2nd week in November in Boston.
It should be interesting trying to sort out what has happened in the various
meetings in the past few months and how that will affect the work of X3J3.
Stay tuned for further developments...
Committee work is a public forum. If anyone would care to express their
opinions at any time, you can submit letters to:
Chairman X3J3
C/O X3 Secretariat/CBEMA
311 First Street, N.W., Suite 500
Washington, DC 20001-2178