ALF ballot resolution meeting, July 9, 2003 Attendees: ---------- Joe Daniels - self Alex Zamfirescu - ASC Kim Nguyen - Tera Systems Peter Christensen - Magma Mark Tillinghast - NTS (phone) Wolfgang Roethig - NEC General update -------------- Wolfgang created a document which contains all ballot comments and the responses. We started to fill in the responses. Before submission to Revcom, all negative comments must be refuted or resolved. If modifications are made to the doc based on the resolution, the document with the modifications must be re-circulated to the balloteers. The balloteers have 10 days to respond whether they change their vote based on the modifications. Next Revcom meeting is in September. To meet the deadline for submission of P1603 to Revcom, the resolution group has to finalize the work by August 1. Review and approval of last meeting minutes ------------------------------------------- Minutes were reviewed and approved. Update on editorial comments ---------------------------- We finalized the response to the editorial comments G02, G03, G05, G07, G08, G12, S04, S18, S19a, S21 Update on style manual conformity (Kim Nguyen) ---------------------------------------------- Kim provided detailed notes on style manual conformity and other editorial items (typo, missing reference etc). Group reviewed the list and decided where to take action. Joe will make the corrections to the doc. Update on abstract, keywords and introduction (Wolfgang) -------------------------------------------------------- Wolfgang provided abstract, search keyword list and introduction. Group made changes to the abstract and search keyword list on-line. Group gave the following feedback on introduction: Joe: - strike 2nd sentence on line 14 Alex: - Provide a month in 1999 instead of "summer 1999" - "ASIC Council" needs to be capitalized - Drop drop "an extension" in the sentence "The ASIC Council encouraged..." - Drop vendor names in sentence "Major EDA vendors..." - Reformulate sentence "Although the ownership .. belongs to IEEE" into "IEEE is the owner ..." - Replace "designers" by "tool develpers, integrators, library creators, users" - Position ALF as language as opposed to a 'dumb' format Mark: - Remove last paragraph - Make 2nd last paragraph the 1st paragraph Ideas for writing appealing things abou ALF in the introduction (all): - Goals for ALF - language - extensibility - well-structured - re-use - conciseness - forward-looking - unique description features (vector expr, arithmetic model) - rigorous construction principles Update on comment S22 (Alex) ---------------------------- Alex found no problem with the grammar. To address he issue with the terminology "ALF_statement_termination", the group suggested to flatten the BNF and thereby eliminating this terminology. It is only used in syntax 1 and has no bearing on the production of the grammar. Current formulation of syntax 1: ALF_statement ::= ALF_type [ [ index ] ALF_name [ index ] ] [ '=' ALF_value ] ALF_statement_termination ALF_statement_termination ::= ';' | '{' { ALF_value | ':' | ';' } '}' | '{' { ALF_statement } '}' Proposed change: ALF_statement ::= ALF_type [ [ index ] ALF_name [ index ] ] [ '=' ALF_value ] ';' | ALF_type [ [ index ] ALF_name [ index ] ] [ '=' ALF_value ] '{' { ALF_value | ':' | ';' } '}' | ALF_type [ [ index ] ALF_name [ index ] ] [ '=' ALF_value ] '{' { ALF_statement } '}' Update on comment G04, S16 (John Williams) and discussion (all) --------------------------------------------------------------- Group reviewed John Williams' updated comment list ith ranking. Since John Williams was not present, discussion was defered. Update on conformance clause comment (Wolfgang) and discussion (all) -------------------------------------------------------------------- Group reached the following consensus: In terms of grammar, ALF is not a "pick and chose" standard. At least, an ALF parser needs to comprehend the full grammar and build datastructures accessible by software, in order to be ALF-compliant. An application is ALF-compliant when ALF is used within an application scope. An application that reads ALF data but does nothing meaningful with it cannot claim to be ALF-compliant. Detailing the meaningful set of data for each application is beyond the scope of this standard. It is potential work for a future standard. A similar approach has been taken with VHDL. VHDL is a general-purpose language, and associated standards (VITAL etc) have been developed to narrow down a particular application. ALF is meant to be a technology enabler. Therefore this standard should stake out the application domains without being too restrictive or exclusive. A critique has been that the application domains described in section 1.2 are all qualified with "can", not a single "shall". Therefore Wolfgang proposed to write one paragraph at the beginning of 1.2 to reflect the consensus of the discussion and thereby address the issue raised in the comments G04, S16 by John Williams and by Mark Tillinghast. Mark felt that this would be a step in the right direction. The group will review the proposed paragraph and decide whether or not to include it. Other ----- During review of the document, Mark made notes for corrections, which he sent to Wolfgang. Next meetings ------------- Wed July 16, 2PM to 4PM phone conference Wed July 20, 2PM to 5PM face-to-face meeting Action items ------------ A.I. Edit introduction for contents and stage new version - Wolfgang A.I. Review introduction and give feedback - all A.I. Edit participant list, mention chair, remove affiliation - Joe Daniels A.I. Edit all sections for conformance to style manual, according to the notes taken - Joe Daniels A.I. Update balloting comment document - Wolfgang A.I. Make all notes and further correspondance on the subject available to resolution group - Wolfgang A.I. Write one paragraph for proposed insertion at the beginning of section 1.2 - Wolfgang A.I. Review paragraph and give feedback - all