From janshep@watson.ibm.com Fri Sep 15 23:31:49 1995 Received: from watson.ibm.com (watson.ibm.com [129.34.139.4]) by dkuug.dk (8.6.12/8.6.12) with SMTP id XAA22582 for ; Fri, 15 Sep 1995 23:31:46 +0200 Message-Id: <199509152131.XAA22582@dkuug.dk> Received: from YKTVMV by watson.ibm.com (IBM VM SMTP V2R3) with BSMTP id 4379; Fri, 15 Sep 95 17:31:29 EDT Date: Fri, 15 Sep 95 16:57:15 EDT From: "Janice C. Shepherd ((914) 784-6313)" X-Addr: J1-K10, Hawthorne I tieline 863 To: sc22wg5@dkuug.dk Subject: defect items 175 and 179 John's comments included: >175 N Reid > I see no need for this to be a part of Corrigendum 3. Has anyone > misinterpreted this term? It is perfectly proper as part of the > Fortran 95 rewrite, however. > > >179 N Reid > This is an unnecessary technical change to Fortran 90, as > explained by Malcolm Cohen. One point to keep in mind about defect 179 "DO variable with POINTER attribute" is that the current answer to defect 179 makes Fortran 90 consistent with Fortran 95. If you don't think the current answer describes what you want to see happen in any version of the standard then you should be sure that your public comment (through whatever means is appropriate for your location (it is too late for public comments in the USA)) reflects this dislike of the restriction in Fortran 95 too. I think the edit in defect 175 'What is a "constant specification expression"?' is appropriate for Corrigendum 3, as the current text is ambiguous as to whether expressions such as "I - I + 1" are constant specification expressions or not. While this example may not be realistic, I can see someone having a named constant IC that sometimes has the value 0, then wondering if "IC * I + 10" is a constant specification expression or not when IC = 0. The answer given in defect 175 is that neither of these expressions are constant specification expressions. I think it is important to be clear on this in Fortran 90. Janice