From martin@ocfmail.ocf.llnl.gov Fri Dec 11 04:54:34 1992 Received: from ocfmail.ocf.llnl.gov by dkuug.dk with SMTP id AA12821 (5.65c8/IDA-1.4.4j for ); Fri, 11 Dec 1992 21:54:35 +0100 Received: by ocfmail.ocf.llnl.gov (4.1/SMI-4.0) id AA27335; Fri, 11 Dec 92 12:54:34 PST Date: Fri, 11 Dec 92 12:54:34 PST From: martin@ocfmail.ocf.llnl.gov (Jeanne Martin) Message-Id: <9212112054.AA27335@ocfmail.ocf.llnl.gov> To: SC22WG5@dkuug.dk Subject: Optional or Not? X-Charset: ASCII X-Char-Esc: 29 To: WG5 From: Jeanne Martin, Convenor Subject: The question of optionality for Varying Length Character Strings With respect to the ongoing CD ballot on the varying length character string standard, I have had three separate inquiries regarding its "optionality". I think we all agree that our original intention was that this standard be a [pick one: {companion, collateral, auxiliary}] standard to ISO/IEC 1539:1991 and that it be optional. That concept does not seem to have occurred to the preparers of our Directives. Also, the ISO Directives we were operating under prior to 1990 mention "Addenda" as a near synonym for "Amendments". The JTC1 Directives we are operating under now have dropped the term "Addenda" and use only "Amendments". An amendment is clearly an extension to a standard that is generally incorporated in the next version of the standard. ______________________________________________________________________________ Here is a copy of our original request to SC22: ______________________________________________________________________________ ISO/IEC JTC1/SC22 Working Group 5 Proposed Project Subdivision for a Fortran Varying Character Module Discussion There have been numerous requests to add a varying character capability to the draft proposed Fortran standard. At its meeting in Paris in 1988, WG5 determined that this need could best be met by a standard module as an addendum to the forthcoming standard revision rather than by adding a varying character feature to the language. The following resolution was passed: P7 - Varying Character Module That WG5 requests the German member body to prepare a proposal for a Fortran module for varying character and the WG5 Convenor subsequently to request SC22 to split the work item to allow standardization of the module. At the 1989 WG5 meeting in Ispra, a delegate from the German member body presented a prototype of such a module. He also requested a minor change in the draft revision which could be made following the SC22 ballot on the second draft. The prototype (See WG5-N418 attached) adequately demonstrated the feasibility of such a module, and the following resolution was adopted at the Ispra meeting: I9 - Varying Character Module That WG5 requests SC22 to subdivide Project JTC1.22.02 Fortran into JTC1.22.02.1 Revision of ISO1539 JTC1.22.02.2 Varying Character Module As a first draft for Project JTC1.22.02.2, WG5-N418 can be used. In adition to providing a standard interface for a varying character capability, this module would serve as an example to others of the use of the new language extension mechanisms defined in the current draft proposed revision of DP1539. WG5 wishes such a module to be standardized as an addendum to the forthcoming revision of DP1539. Proposed Resolutions Resolution 1 ISO/IEC JTC1/SC22 subdivides its work item JTC1.22.02 as follows to provide a Fortran varying character module as an addendum to the revision of ISO1539: JTC1.22.02.1 Revision of ISO1539 JTC1.22.02.2 Varying Character Module Resolution 2 ISO/IEC JTC1/SC22 requests that the German member body provide working drafts of an addendum to the revision of ISO1539 for a varying character module. ______________________________________________________________________________ Here is the SC22 Resolution adopted at the Berlin SC22 Meeting, Sept. 26-29, 1989 (WG5-N442) ______________________________________________________________________________ Resolution 132: Subdivision of work item JTC1.22.02 - Fortran ISO/IEC JTC1/SC22 - authorizes the subdivision of work item JTC1.22.02 - Fortran as follows: JTC1.22.02.01 - Revision of ISO 1539 - Fortran; and JTC1.22.02.02 - Variable length character string module; - notes that SC22 documents relevant to JTC1.22.02.02 are N680 and N723; and - requests the German Federal Republic member body to provide a project editor for JTC1.22.02.02. ______________________________________________________________________________ At this point I can think of three possible directions we could take to achieve our original objective: 1. Retain our current approach, but put the word "optional" on the title page and put something like the following in the Introduction - "This standard is optional. A processor conforming to ISO/IEC 1539:1991 is not required to provide facilities for the manipulation of character strings of dynamically variable length, but if it does, it should provide them in a manner that conforms to this standard." This is made explicit in the Scope (more than once) but, in my opinion, that is not early or obvious enough. All instances of the term "collateral" should be excised. This term does not have a definition in the Directives. JTC1 does not have procedures for optional standards and is unlikely to devise any for our needs alone. By taking this approach we would be using the existing procedures, but making them work to our ends. 2. Recast the draft as a Technical Report - Type 3 instead of a standard. 3. Recast the draft as a binding to a varying length character string facility. A drawback with this approach is that to have a binding, there may have to be an existing language independent standard to bind to. The direction we choose to take will be discussed and settled at our next meeting in Germany. A member body ballot may, of course, indicate a preferred approach, but the main purpose of the ballot is to flush out technical issues such as: Are the right facilities provided? Is I/O handled correctly? Are the names of the module and data type appropriate?