From malcolm@brackley.nag.co.uk Fri Mar 31 19:23:45 2000 Received: from brackley.nag.co.uk ([62.232.54.10]) by dkuug.dk (8.9.2/8.9.2) with ESMTP id TAA31160 for ; Fri, 31 Mar 2000 19:23:45 +0200 (CEST) (envelope-from malcolm@brackley.nag.co.uk) Received: (from malcolm@localhost) by brackley.nag.co.uk (8.9.2/8.9.2) id SAA78342 for SC22WG5@dkuug.dk; Fri, 31 Mar 2000 18:22:53 +0100 (BST) (envelope-from malcolm) From: Malcolm Cohen Message-Id: <200003311722.SAA78342@brackley.nag.co.uk> Subject: (SC22WG5.1765) Second revision of proposal for Interpretation 001 To: SC22WG5@dkuug.dk Date: Fri, 31 Mar 2000 18:22:53 +0100 (BST) X-Mailer: ELM [version 2.4ME+ PL43 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Craig T. Dedo wrote: > 1. Are you taking ownership of this particular interpretation? From the text >in your introductory paragraph it appears that you are. No-one "owns" an interpretation, unless it be jointly - the original submitter - J3/INTERP subgroup as a whole. After all, we're not doing MTE-style language design here, or constructing vast wordy edifices, we are trying to come to a consensus about what the standard says or was intended to say. I don't see how some form of "ownership" is going to advance these goals, nor do I see a mention of "ownership" in our procedures. From John's text I would take it rather that he is intending to submit his draft answer as a J3 paper for the next meeting to take action on. And very welcome it is too. Just my opinions of course. > 2. What about the KIND type parameter of the index variable? Is it always >default KIND? Is the KIND type parameter derived from the values of the other >variables in the statement? Is it always of the KIND with the largest possible >value? You quoted the ANSWER which quoted text from 14.1.3 which answered all these questions for implied DO. I shan't include it here, but you can find it in the F95 standard at [280:40-44]. For FORALL, see 7.5.4.1 "The FORALL Construct", particularly [115:25-28]. > 3. If there is a variable declared in the scoping unit with the same name and >it is suitable for use as an index variable: > a. Is the declared variable used, or > b. Is the index variable different from the declared variable? > Example: > INTEGER(KIND=8) :: I ! On this processor, KIND=8 specifies 64-bit >binary arithmetic. > REAL, DIMENSION :: A(10,10), X(10) > ! Processing during which I is assigned the value of 10 billion (10**9) >and has that value immediately prior to the following statement. > FORALL (I=1:10) A(I,I) = X(I) >After the execution of the FORALL statement, does I have the value 10 or 10 >billion? Or, is I undefined? >FWIW, I would prefer option b, so that I after the FORALL statement has the value >10 billion. Please read 7.5.4.1, in particular [115:25] where it says "An in a has a scope of the construct (14.1.3)" That really says it all, but the referenced 14.1.3 "Statement and construct entities" goes into gory detail, in particular see [281:21-24]. > 4. Should we have a straw vote, either by e-mail or at the next J3 meeting on >the questions in item 2 and 3? No. Cheers, -- ...........................Malcolm Cohen, NAG Ltd., Oxford, U.K. (malcolm@nag.co.uk)