From owner-sc22wg5@open-std.org Sat Dec 11 12:03:13 2004 Return-Path: X-Original-To: sc22wg5-domo1 Delivered-To: sc22wg5-domo1@open-std.org Received: by open-std.org (Postfix, from userid 521) id ADB541ED69; Sat, 11 Dec 2004 12:03:13 +0100 (CET) X-Original-To: sc22wg5@open-std.org Delivered-To: sc22wg5@open-std.org Received: from rap.rap.dk (213.237.47.228.adsl.vbr.worldonline.dk [213.237.47.228]) by open-std.org (Postfix) with ESMTP id 52B4A876B for ; Sat, 11 Dec 2004 12:03:07 +0100 (CET) Received: by rap.rap.dk (Postfix, from userid 500) id 627DB2D9DE; Sat, 11 Dec 2004 12:03:03 +0100 (CET) X-Original-To: sc22wg5@open-std.org Received: from dkuug.dk (ptah.dkuug.dk [195.215.30.66]) by open-std.org (Postfix) with ESMTP id 343B812E34 for ; Sat, 11 Dec 2004 05:21:10 +0100 (CET) Received: from avradionet.com ([67.99.246.240]) by dkuug.dk (8.12.10/8.9.2) with ESMTP id iBB4I7jH012710 for ; Sat, 11 Dec 2004 05:18:15 +0100 (CET) (envelope-from maine@summertriangle.net) Received: from [192.168.1.2] ([67.99.246.111]) by avradionet.com (avradionet.com) (MDaemon.PRO.v7.2.0.R) with ESMTP id md50000400643.msg for ; Fri, 10 Dec 2004 20:19:27 -0800 Date: Fri, 10 Dec 2004 20:15:24 -0800 From: Richard Maine To: sc22wg5@dkuug.dk Subject: RE: WG5 letter ballot on interpretations Message-ID: X-Mailer: Mulberry/3.1.6 (Win32) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-Spam-Processed: avradionet.com, Fri, 10 Dec 2004 20:19:27 -0800 (not processed: message from trusted or authenticated source) X-MDRemoteIP: 67.99.246.111 X-Return-Path: maine@summertriangle.net X-MDaemon-Deliver-To: sc22wg5@dkuug.dk X-Spam-Score: 0 () Sender: owner-sc22wg5@open-std.org Precedence: bulk (I'm not entirely sure whether the list will accept email from my home address. If it doesn't go through, I'll resend from work Monday). A general comment - I found this set of interps so confused that I wish I could vote "take the whole thing back and try again", but that doesn't appear to be an option. The global problem is confusion as to what standard these interps are about. The ballot title says that these are f2003 interps. However, many of them seem to be f95 interps instead of f2003 ones. Many of the f95 interps are about issues that also apply to f2003, but the citations are all to f95. Others don't seem to have anything to do with f2003 or give answers that are no longer correct for f2003. I encourage everyone to actually read these interps and see if they make sense for f2003 and if it is clear what standard they are referencing. While I appreciate the desire to get the old f90/f95 interps done, particularly where they still apply to f2003, I don't think we do anyone any favors by publishing "clarifications" where the reader is given no hint that the citations refer to f95 or f90 instead of f2003. Having f90 in the interp number is not sufficient to define the references as being to f90 (and such an inference would not be consistently correct anyway). I might think that this ballot was for f95 interpretations, but some of the items are about features new to f2003. The following Fortran 2003 interpretations are being balloted: Yes No Number Title --- -N- 000004 Value returned by MAXVAL/MINVAL --- -N- 000006 Character length specification of a function result -Y- --- 000008 Optional arguments to MAX/MIN --- -N- 000017 Characteristics of an array function result --- -N- 000023 Termination of the previous record by a WRITE statement --- -N- 000030 Ordering requirements on definition of specification functions --- -N- 000031 Association of pointer function result with INTENT(OUT) dummy argument (subsumed by 000074) --- -N- 000068 Asterisks as I/O units --- -N- 000074 TARGET dummy arguments and POINTER expressions --- -N- 000078 Resolving generic procedure references --- -N- 000096 End-of-record and PAD --- -N- 000098 Are dummy functions returning assumed-length character legal? (duplicate of 000006) --- -N- 000102 mask-expr evaluated only once --- -N- 000103 Derived type name DOUBLEPRECISION --- -N- 000104 Representation method of result of REAL --- -N- F90/000049 Characteristics of function results --- -N- F90/000070 Characteristics specified by interface bodies --- -N- F90/000096 Definition of "Declaration" --- -N- F90/000140 TARGET attribute for a derived-type object with a pointer component --- -N- F90/000180 Unambiguous generic references --- -N- F90/000206 Collating sequence inconsistencies --- -N- F90/000207 Integer bit-model inconsistency --- -N- F90/000208 nonadvancing output followed by list directed output -Y- --- F90/000210 nonadvancing write followed by list directed write --- -N- JP-24 The bnf term shared-term-do-construct -Y- --- F03/0001 Generic type-bound procedures -Y- --- F03/0002 Component value for pointer components -Y- --- F03/0003 Referencing deferred bindings {subsumed by F03/0004} -Y- --- F03/0004 Type-bound procedures and undefined association status -Y- --- F03/0005 Argument association and the TARGET attribute -Y- --- F03/0006 Intrinsic assignment and allocatable components -Y- --- F03/0007 Finalization of structure constructors in specifications -Y- --- F03/0009 VALUE attribute for passed-object dummy arguments -Y- --- F03/0010 Unlimited polymorphic pointer/allocatable dummy arguments -Y- --- F03/0011 Allocating objects of abstract types -Y- --- F03/0013 VALUE attribute for polymorphic dummy arguments -Y- --- F03/0014 Automatic arrays in interface bodies -Y- --- F03/0015 TARGET attribute for associate names -Y- --- F03/0016 Invoking type-bound procedures via array objects Redasons for "NO" votes. 004 The answer cites 13.14.39, which does not exist in f2003. This citation should either be changed or explicitly documented as referring to f95 (which I presume it does - I didn't check). The technical content is otherwise ok. 006 The interp explicitly discusses f95, without even mentioning that the same principles apply to f2003. The technical content is otherwise ok. 017 The interp explicitly discusses f95, without even mentioning whether the same principles apply to f2003. I'm not sure whether it is relevant or not. 023 The interp references 9.1.3.2, which doesn't exist in f2003. There is no indication that this might be a reference to f95. Likewise for other section and page/line citations. 030 The interp references 7.1.6.1, which doesn't exist in f2003. There is no indication that this might be a reference to f95. Likewise for other section and page/line citations. Notes 7.14 and 7.16 exist, but are implausible places for the edits. 031 This appears to be an f95 interp. If it is worth listing at all as an f2003 one, even subsumed, it should at least mention that its citations are to f95 instead of f2003. Also, I voted against interp 74, and I don't think we can count this one as being closed by interp 74 if interp 74 isn't closed. 068 This interp is explicitly about f95. Furthermore, in this case, the answer is distinctly different in f2003. I don't see how we can claim this to be an f2003 interp as is. Just changing the citations won't fix this. 074 This interp references sections that exist in f2003, but do not contain the cited material. There is no indication that these might be references to f95. 078 The interp references sections that do not exist in f2003. There is no indication that these might be references to f95. 096 The interp references sections that do not exist in f2003. There is no indication that these might be references to f95. There are f2003 edits, but those edits cite 03-007r2, which is not the latest f2003 internal document. It looks to me like the pages and lines are still correct for the latest internal document, but the citation should be updated. 098 This appears to be an f95 interp. If it is worth listing at all as an f2003 one, even as a dup, it should at least mention that its citations are to f95 instead of f2003. Also, I voted against interp 6. 102 The cited line numbers don't exist on that page in f2003. There is no indication that these might be references to f95. 103 This makes no sense at all as an f2003 question. It is a question about wording that doesn't exist in f2003. The suggested "clarification" is exactly what f2003 already says. Also, the only hint that the whole thing is about f95 is the parenthetical acknowledgement that the fix is already in f2003. 104 This interp specifically cites f95 (though you would have to know internal J3 document numbers to realize that). It makes no mention of f2003. f90/000049 The references section explicitly cites f90. There is no mention of whether it still applies to f2003. f90/000070 The references section explicitly cites f90. F2003 is mentioned only as an aside about a "possible edit". If this is actually an f2003 interp, then it needs to be a little more definitive than alluding to a possible edit. f90/000096 The interp is explicitly about f90. There is no mention of whether it still applies to f2003. f90/000140 The interp is explicitly about f90 and f95. Furthermore, the answer is incorrect for f2003. It is no longer true that " The constraints that follow a syntax rule, or a set of syntax rules, are syntactic constraints and apply only to the syntax rules they immediately follow." Constraints include specific indication of what syntax rule(s) they apply to; it is not always the rules that they immediately follow. f90/000180 The citations are clearly not to f2003. There is no mention of what they are to or of whether they still apply to f2003. f90/000206 The citations are explicitly to the f95 cd (not even to the f95 standard). There is no mention of whether they still apply to f2003 (or even to the published f95). f90/000207 The citations are explicitly to the f90. I suppose that the edit must also be - it certainly isn't for f2003. There is no mention of whether they still apply to f2003. f90/000208 The interp references sections that do not exist in f2003. There is no indication that these might be references to f95 or f90 and, if so, whether they still apply to f2003. JP-24 The citations are not to f2003. Only from a parenthetical comment can one deduce that they refer to f95. There is no indication of whether they still apply to f2003. The edit is clearly not an f2003 edit. -- Richard Maine