ISO/IEC JTC1/SC22/WG5/N1454

Subject:  Comments arising from subgroup review of Section 9
To:       WG5
From:     Van Snyder
Date:     2 August 2001

Three areas of concern arose from reviewing Section 9.

Although subgroup has given some attention to each of these concerns,
there has not been time to give them all of the attention that they may
need.  Therefore, the revisions proposed below may not actually be
practical.

1.  At [161:35] it is specified that "the length of a formatted record is
    measured in characters."  At [164:30-45] it is specified that a
    formatted stream file consists of file storage units.  The discussion
    of file storage units in subclause 9.2.4 re-enforces this
    specification, and makes it clear that a file storage unit is not
    necessarily a character, although there is a recommendation that it
    be an 8-bit octet.  This conspiracy appears to result in a
    contradiction if a file that is written (read) by formatted stream
    input/output is read (written) by formatted sequential input/output,
    unless a file storage unit is a character.  This, of course, defeats
    the purpose of file storage units being abstractions.  WG5 wishes that
    J3 verify that there is no problem.

2.  At [174:18] it is not specified what happens if STATUS="SCRATCH" and
    FILE=... are both specified, and the specified file exists.  Because
    of this problem, and the fact that the desired functionality can be
    provided with named files, WG5 prefers that the named scratch file
    facility be deleted.

3.  At[196:21-22], the sentence is not helpful, in that the only
    specifiers to which it applies are UNIT= and FILE=.  Perhaps it should
    be deleted.