J3/04-267r1

Date:    4 Mar 2004
To:      J3
From:    Richard Maine
Subject: Proposed positions on UK FCD comments

Edit subgroup has studied the UK FCD comments and
recommends the following positions on them.

C1.  Array constructor restriction.
     Agree with the idea.  Suggest slight edit improvement.
     Add "in this case" after the semicolon.

C2.  ASYNCHRONOUS and VOLATILE
     The edit on 411 is simillar to one in the US position;
     we slightly prefer the US words.

     The edit on 251 is a simillar case that we missed; we agree.
     See [252:30-34 for support]

     We minorly disagree with adding the xrefs  We don't think
     this an appropriate time for such changes; plus these
     particular ones are misleading because they omit cases.

C3.  VOLATILE
     Ok on the edits on page 201 and 421, which fix errors.

     Ok on the edits on 411 with the suggested improvement
     of "module or internal" -> "internal or module"; that
     avoids possible misparsing.

     No on the new note; we don't think adding new notes now
     is appropriate.  Plus, this one may not be quite the
     full story.

C4.  IMPORT
     May need work.  No particular position established.

C5.  Interoperability
     Ok on the first edit.

     No on the 2 constraints.  Turning requirements into constraints
     is not fixing errors, but could introduce them.

C6.  'Linkage' of procedures and variables
     Agree.  These were probably omissions from a previous change.

C7.  VALUE attribute
     Disagree with deleting the constraint; possible implementation
     problems.   Suggest instead generalizing it to require
     initialization expressions for len type params of all types, but
     drop the special value 1 (which has special meaning only
     for character).

C8.  Binding labels
     Comments are based on the rationale instead of the particular
     edits.  The edit details can be worked depending on the outcome
     of the main questions.

     Rationale item 1 is the same as a proposed US item that we
     already voted against.

     Rationale item 2 we agree with.  This is a simple fix (part,
     but not all of the 403:35 edit).

     Rationale item 3 we think is already implied.  We strongly
     feel that it should not be a constraint.  It might be
     reasonable to state it explicitly as a text requirement;
     we have no strong position either way on that.

     The edits also presume that we want the capability of
     specifying that a bind(c) procedure has no binding label.
     Straw vote whether we support this or not.

C9. Comma before BIND(C)
    We agree that these are errors and need to be fixed.

E1. ISO_FORTRAN_ENV module
    We disagree.  Part of this doesn't even clarify anything.
    The other part is unnecessary.

E2. ISO_C_BINDING module
    We agree.  The draft is in error.

E3. Miscellanea
    Ok.
    Also fix "help" ->"held" on [487:22]


E4. Glossary
    We agree with updating the erroneous entries.

    We slightly prefer not to add new glossary entries now,
    but it is a soft no, there being only two.

E5. Index
    The E5.1 edit is unnecessary wordsmithing in nonnormative text.

    We think this an inappropriate time to be doing index additions.
    Among other things, index entry has high likelihood of adding
    new typos.  However, this is a soft position.