ISO/IEC JTC1/SC22/WG5 N753


informal JLW report of X3J3 meeting 120, Nov 11-15, 1991
--------------------------------------------------------

The three principal objectives of the meeting were to (1) process those 
requests received so far for interpretation of Fortran 90, (2) develop a 
response to the Boeing NO vote on the X3 ballot, and (3) determine the US vote 
on the WG5 ballot regarding moving the varying string proposal to CD status.

There were 15 interpretations "2/3 voted" at meeting 119; these were all 
approved by roll call vote at meeting 120.  In addition, 29 more requests for 
interpretation were processed at meeting 120.  Of these 29, 18 were approved by 
roll call vote, 9 were "2/3 voted", and 2 were voted on but did not receive the 
necessary 18 affirmative votes to be considered "2/3 voted" (the vote on 120-
88a was 15-2, the vote on 120-95 was 17-0).  All 29 will be reaffirmed or 
reconsidered at meeting 121.  There are still two interpretation requests for 
which no action has been taken.  These are 119-49 (assigned to GEN) and 119-
57/120-89 (assigned to PROC); both will be considered at meeting 121.

Early in the meeting a motion was passed (17-8) to "have a goal of handling 
interpretations within one meeting".  The committee worked effectively toward 
this goal during the week, but at the end the general feeling seemed to be that 
the process was being rushed too much.  Therefore the decision was made to 
reaffirm at the next meeting all interpretation actions initiated at this 
meeting.  In addition, though no formal action was taken on this, an apparent 
consensus emerged along the lines that only those requests meeting both of the 
following conditions could be processed at a single meeting:

    (a) the request appears in the premeeting distribution and the
        requestor is invited to the meeting

    (b) a proposed interpretation is distributed (email OK) two weeks   
        before the meeting

In all other cases the final role call action on an interpretation should be 
delayed to the following meeting.

It was decided that all requests for interpretation and the disposition of 
those requests will be recorded in an X3J3 standing document (S20), which will 
be maintained by Andrew Tait, published in Fortran Forum, and available to all 
interpretation requestors and others.

The response to the Boeing NO vote (120-16b) was approved by a role call vote 
of 23 yes, 0 no, and 1 abstention (Philips).

The US TAG voted 15-0-1 that the US vote NO on the N743 (varying string module) 
CD ballot.  A list of seven specific technical changes are required to change 
this vote to YES.  Generally, these seven involve tightening up the 
functionality specified by the proposed standard and none involve major 
changes.  A few editorial suggestions are also included.

The surprise development of the meeting was the discussion and straw vote (21-0-
1) on Tuesday that a near-term "maintenance" revision of Fortran 90 be produced 
to incorporate the interpretation actions.  On Friday the US TAG adopted a 
resolution (14-2) that a "maintenance" revision, limited to clarifications, 
edits, corrections, and the results of interpretations, be published in 
approximately the 1995 time frame.  A companion TAG resolution (9-6) 
recommended that the development phase of a concurrent evolutionary revision be 
completed in approximately the 1998 time frame.  

In the final action of this meeting, X3J3 voted (18-2) to have a letter ballot 
establishing a treasurer and a meeting fee.  This should be done before the 
next meeting, which is scheduled for May 25-29, 1992, in Terre Haute (Mallory 
North host).

Jerry Wagener
91-11-18