From maine@altair.dfrc.nasa.gov Tue May 21 19:41:30 1996 Received: from altair.dfrc.nasa.gov ([130.134.34.72]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id TAA04696 for ; Tue, 21 May 1996 19:41:28 +0200 Received: by altair.dfrc.nasa.gov (SMI-8.6/SMI-SVR4) id KAA21334; Tue, 21 May 1996 10:42:01 -0700 Date: Tue, 21 May 1996 10:42:01 -0700 Message-Id: <199605211742.KAA21334@altair.dfrc.nasa.gov> From: Richard Maine To: sc22wg5@dkuug.dk cc: J.S.Morgan@liverpool.ac.uk Subject: Liaison Report on Parameterized Derived Types Enclosed is a copy of the X3J3 liaison report on the proposed TR on parameterized derived types. This report was approved as paper 96-109R1 at X3J3 meeting 137. Steve, this is identical to the version (with your change) that you saw earlier in the meeting, except for the correction of a few of my spelling errors. -- Rich Maine -------------------------------------------------------------- X3J3/96-109R1 To: X3J3, WG5, and Steve Morgan From: Richard Maine Subject: Liaison Report on Parameterized Derived Types At X3J3 meeting 137, Steve Morgan made a presentation on the proposed TR on parameterized derived types. Substantial discussion ensued on the technical concerns that had been previously raised by X3J3 in paper X3J3/95-0288R1. There appears to be general agreement on the approach to handling several of the concerns. In some cases X3J3 appeared satisfied that an issue was adequately answered; the question of "*" versus ":" for dummy nonkind parameters was one such issue. In other cases, Steve Morgan agreed with the concern and with an approach to addressing it; the lack of edits in section 4 to define the syntax of appropriate constructors was one such issue. A general agreement does not yet appear to have developed on the question of whether inquiries are best handled by a syntax like an inquiry function or by a syntax like component selection. The current proposal uses an inquiry function syntax. There was substantial sentiment for a syntax like component selection. Several other areas were also the subject of debate, but most of these seem to tie closely to the issue of whether or not derived type parameters look somewhat like components. For example, the question of namespace "pollution" and the effects of rename become moot if a component selection syntax is used instead of an inquiry function syntax. Similarly, the syntax for component definition allows no possibility of implicit typing, so if the type parameter specification uses similar syntax, the question of implicit typing also becomes moot. However, using a component selection syntax would be potentially confusing in that such "components" could be referenced, but never set. The main conclusion drawn from this discussion was that further work was needed to reach sufficient consensus on the best technical approach. Steve Morgan felt that there was still sufficient time to produce an alternative TR (taking into account the discussions at the meeting) for presentation at the WG5 meeting in Dresden. -- Richard Maine maine@altair.dfrc.nasa.gov