ALF ballot review meeting, July 1, 2003 Participants: John Michael Williams, Markanix Alex Zamfirescu, ASC Inc. Mark Tillinghast, NTS Corp. (by phone) Joe Daniels, IEEE (by phone) Kim Nguyen, Tera Systems Peter Christensen, Magma Arun Balakrishnan, NEC Minutes taken by Wolfgang Roethig, NEC Review of ballot result ----------------------- 37 out of 47 votes received 32 affirmative votes 2 negative votes 3 abstention votes Details see the following documentation, which was distributed in the meeting and staged on the web (www.eda.org/alf/homepage/ballot.zip): P1603_bal-sum.final - details of ballot results ballot.hanna.doc - affirmative vote w comment ballot.jwill.doc - negative vote w comment ballot.shanka.txt - affirmative vote w comment ballot.barrow.txt - administrative comment ballot.kmchow.txt - affirmative vote w comment ballot.mtillinghast.txt - negative vote w comment Explanation of procedure ------------------------ #1 Fill in the introductory sections: abstract, search keywords, introduction, WG member list (mandatory) #2 Respond to and resolve comments associated w negative vote (mandatory) #3 Respond to comments associated w other vote (optional) #4 submit work to Revcom for ratification within 1 month (contingent on #1 and #2) Recirculate doc to ballot group after completing #1 and #2. If #1 and #2 deemed satisfactory by ballot constituency, then std can be ratified by RevCom. Discussion of negative comments ------------------------------- ballot.jwill.doc, G04, S16 as well as ballot.mtillinghast.txt revolve around the relevance of the "overview" as a normative clause and the absense of a compliance clause. The overview mainly paraphrases the PAR and application flows. The language in the text says "an application can use ALF ..." Mark requested a compliance clause, for example as suggested by John Williams in his comment S16. Alex recommended against a compliance clause. A clause that stated something like "a design flow compliant w the std shall use the std" would be redundant. For any application of any std it is understood that compliance can only be claimed if the std is used. On the other hand, ALF can be used by many applications in many different ways, and specifying a detailed compliance test is material enough for another std. Precedence: VHDL is a general language std, complementary stds exist for synthesizable VHDL, std math package etc. The WG defended the usage of "can" in a normative clause. Normative clause does not mean, everything is mandatory. Normative clauses make the distinction between mandatory and optional parts of a standard through the usage of "shall" and "can". Proposed resolution: Explain the rationale of the ALF standard and the conformance/compliance aspect in the introduction, which has to be supplied anyway. Resolution of negative comments ------------------------------- G01, G03, G07, G08, G10 These comments are all about usage of fonts and types. Joe will double check that the used fonts and types are compliant with the IEEE style manual. Any non-compliant fonts and types can be changed by the editor, without requiring re-ballot. This resolution is deemed acceptable. S01, S02, S03 These comments are about missing abstract, introduction, participant list. the missing parts will be supplied by the next meeting. They are not subject to ballot. This resolution is deemed acceptable. S22 The metalanguage was mis-understood. The concern was that the last statement within a block, i.e. enclosed by { and }, needs no terminating semicolon. However, this is not the case. Alex will confirm with aSC's ALF parser that the metalanguage indeed defines a rigorous requirement for ALF-statement termination by semicolon. This resolution is deemed acceptable, pending proof of correctness. Minor corrections to doc ------------------------ Figure 1 should have a dashed connection between "ALF" and "layout editor". This, BTW, resolves John Williams' non-pertinent comment S12. "mantisse" should be "mantissa" throughout the doc. Action items ------------ Supply the missing sections according to style manual - Wolfgang (This supercedes the originally proposed AI of checking the doc against style manual) Style manual is available at standards.ieee.org/guides/style/index.html Proof correctness of meta-language as defined in syntax 1 - Alex check all the editorial items, ie., G01, G03, G07, G08, G10 - Joe Next meeting ------------ Wed July 9, same time (2PM), same place (NEC)