Phone Conference P1450.1 Working Doc Subgroup Thurs Apr 11, 10:00 am PDT --------------------------------------------- Attendees: ---------- Tony Taylor Greg Maston Peter Wohl Doug Sprague Jim Teisher David Kellerman Documents --------- p1450.1 D13 draft March 21. D13 review-resolution document Apr 5. overview section proposal, pdf email from Tony. Successes --------- Tony closed out the D12 draft and resolution doc, and distributed the documents to all first-round reviewers. With this closure, the Working Group can move to a *new* review-resolution document for D13 issues (and ultimately resuling in a D14 draft). Agenda ------ Discussed review of clarification doc for dot0. Deferred this to the next meeting as Greg sent a reply to an inquiry from David in regards one of the clarifications, and Doug will collect some IBM responses as well. The agenda was established with a review of the D13 review-resolution doc and selection of issues to address. Issues ------ TT-1/GR-1: Started with a review of the dot1 overview proposal by Tony. This section has resulting in some email discussions, with the expectation of oncoming revisions to this text. Draft D14 will contain this overview, subject (as is all the document) to continued review. GR-2: clash on extensions. Response: All STIL dot Working Groups coordinate on all extensions to the base language; all extensions ARE consistent for the application of each dotted effort. No change identified. GR-3: persnicky expressions. Response: single-quotes are consistently used to identify expressions in the language. No change identified. GR-4/GR-10: UserKeywords in local contexts generates the need for many repeated definitions. Working Group expects two usage contexts: global user definitions, or local user definitions. A specific application (that has need of Userkeywords) would operate in one of these two contexts, and if the local context was at a level that required repeated definitions then the tool should consider moving the definition of that context to a higher level. See example at the end of these minutes. The notion of a "global definition that has local consequences" is excessive support for this construct which is intented to have MINIMAL application. The Working Group discussed the reference of XML and notions of a general language. The context of the Working Group is that STIL is NOT intended to be a general language, and that extending the language toward generality is not the goal of the efforts. However, the Working Group understands that *something* might be done here but does not understand the requirement in order to address it. GR-5: "contents of the spec identify a scope that can be disagreed with". While this effort (under the umbrella of a standards definition) must support sufficient definition to be used in the industry, the definition of needs and requirements is determined by participation in the effort. If specific issues can be identified requiring enhancements to constructs defined, then the Working Group is always overwilling to augment the existing definitions. The definitions currently present represent the Working Group's efforts at defining meaningful usable constructs. GR-6: Move Glossary to clause 3. All definitions in clause 3 become part of the IEEE list of standard definitions, and the previous ballot process of 1450.0 resulted in moving concepts that are local to a spec, into an Annex. If there are specific recommendations of definitions that ought to be part of the IEEE context then we're more than willing to move them into clause 3. GR-7: "integers are not signalgroups." This issue is well-understood by the Working Group, which has moved these constructs all around various areas of the language. While the concern is appreciated, the lack of a recommendation makes the Working Group unable to address this concern. Tony will follow-up with an email to Gordon on GR-4, GR-6, GR-7, in an effort to provide more information to the Working Group toward recommendations to satisfy these issues. Next Meeting ------------ in two weeks; Apr 25, 2002. Meeting was ajourned at 11:00 PDT. ------------------------------- Example of "redundant userkeyword definitions". Suppose a userkeyword was meant to be applied only INSIDE a Vector statement, for example: Pattern a { V { UserKeywords UseMode; allsigs=00000XXXX; UseMode 1; } V { UserKeywords UseMode; a=1; b=X; UseMode 1; } V { UserKeywords UseMode; a=2; UseMode 2; } V { UserKeywords UseMode; allsigs=0110HHHLL; UseMode 2; } V { UserKeywords UseMode; allsigs=0000HHHHH; UseMode 2; } } This is an expensive definition. However, by moving this definition up to inside the Pattern block, this redundancy is eliminated: Pattern a { UserKeywords UseMode; V { allsigs=00000XXXX; UseMode 1; } V { a=1; b=X; UseMode 1; } V { a=2; UseMode 2; } V { allsigs=0110HHHLL; UseMode 2; } V { allsigs=0000HHHHH; UseMode 2; } } However, there is the negative impact that this userkeyword now becomes potentially accessed under the OTHER statements of this block, for example in a C{} statement. The Working Group understands this limitation. The UserKeywords construct is of very limited application, and has required extensive justification in previous balloting cycles to support limited functionality. Augmenting this construct implies an importance to this construct that it is not intended to have, and the Working Group is hesitant to augment this construct further at this time (changes in membership of the Working Group always opens this decision up for review, however).