Subject: Notes on the ATG2 meeting in Washington
For use in tomorrow's Atlantic TC call. ================================================================== NOTES ON THE ATG2 MEETING IN WASHINGTON The ATG2 meeting took place the week of 15 January 2007 at SAP offices in Washington, DC. Organizations represented included: SAP Oracle Sun Microsystems U.S. Navy OAGI STAR AIAG X12 (informal observer) OASIS (UBL) GS1 RosettaNet ACORD UBL was represented by UBL chair Jon Bosak, who attended only the plenaries and the work sessions devoted to NDR revision. ================================================================== UBL INPUTS UBL submitted the following methodology papers: THE ARGUMENTS FOR EXTENSIBLE CONTENT.pdf UBL-codelist-methodology-0.8-D2-20070102-2240z.zip gkholman-ubl-modeling-0.4.zip And the latest UBL NDRs: prd-UBL-NDR-2.0.pdf And a copy of the UBL 2.0 Standard: os-UBL-2.0.zip ================================================================== SALIENT ITEMS FROM THE OPENING PLENARY - ATG2 has no mailing list of its own but rather uses the ATG mailing list. The ATG list is managed by the UN; there is no publicly visible archive, and all requests to be added to the list must go through the ATG chair. The ATG web site is hosted by CEN. - CCTS version 2.01 is the basis for NDR 2. CCTS version 3 should be available in February. NDR 3 will be based on CCTS 3. The datatype work that used to be part of CCTS is being extracted out into the DT project in ATG and will be used in NDR 3. - The ATG chair gave as his informal opinion that UBL 2.0 is still "in the ballpark" of complete conformance to CCTS3. - The UBL chair stated UBL's understanding that UBL and UN/CEFACT have agreed that UBL will contribute its work to UN/CEFACT. He observed that UBL is unique in this respect among the other efforts represented at the meeting and stated that UBL should not be considered in the same category as the other SDOs for purposes of the ATG NDR work. - The UBL chair said that he understood why UN/CEFACT would wish to benefit from the implementation experience of the SDOs represented in the meeting in constructing its own next generation of NDRs, but did not understand the benefit to be gained by the NDOs in adopting a common set of NDRs. Members present replied that user communities were in favor of common NDRs and that there was some evidence that convergence on common NDRs would promote harmonization beyond what was achievable with CCTS convergence alone. - The UBL chair suggested adding the following business requirement: Keep NDR changes incompatible with the UBL 2.0 schemas to a minimum to facilitate the migration of the UBL installed base to future UN/CEFACT offerings. This suggestion was rejected. ================================================================== CODE LISTS Following a discussion largely driven by UBL and GS1 with the assistance of Marty Burns, the following were accepted as requirements for the code list work (these are from my notes and must be checked against the minutes of the meeting when those become available): - A standard xml form of representation for the publication of code lists (e.g., genericode, now under development by the OASIS Code List Representation TC) - The ability to subset or extend code lists without requiring changes to the standard schemas - The ability to associate different versions of a code list with different locations in the instance document - The ability to associate conditional logic with code values - A standard way to specify all this in individual trading partner agreements - Support for code values in both elements and attributes - Support for multiple lists of equivalent codes (covered by genericode) - Support for code list metadata (covered by genericode) OAGi observed that they have been using a second XSLT validation pass since 2002. ================================================================== OTHER NDR ISSUES ATG2 briefly reviewed most of the current ATG NDRs to note items that needed work. No attempt was made to take notes on these; UBL NDR editors will need to review the issues matrix when that is made available. ================================================================== RELEVANT ACTION ITEMS ACTION: Every SDO to contribute two slides on derivation [= customization or extension] by the beginning of February for use during the Dublin UN/CEFACT Forum in March: Slide 1: The business case for derivation/extension Slide 2: The current approach used by the SDO in meeting the requirements Slides to be sent to Garret Minakawa (email@example.com), Don Allen (firstname.lastname@example.org), and Serge Cayron (email@example.com). ACTION: ACORD and OAGI will write a paper on derivation at the model and schema levels taking UBL inputs into consideration. ACTION: All SDOs to review the current ATG NDR "hybrid" approach for local/global in preparation for further discussion. This approach was summed up as "declare all associations as aggregations." ACTION: ATG2 chair to "distribute ATG2 minutes elaborating on the hybrid approach." ACTION: UBL NDR editors to review the new CCTS typology, to be forwarded by the ATG chair. ================================================================== MISCELLANEOUS - The UBL chair suggested that ATG2 use the NDR authoring format created by the Quality of Design team at NIST. - ATG message 249 from Garrett looks like something that PLB needs to review. - ATG2 will be working again in Dublin at the UN/CEFACT Forum meeting. - The next interim ATG2 meeting is tentatively set for the week of 25 June 2007 at GS1 offices in Princeton (NJ). - One-hour NDR working calls will be held every Thursday at 16:00 UTC (11:00 EST) beginning the week of 29 January [my notes say "next week," which would be the week of 22 January, but I don't think that this is correct]. Core Data Type calls will be held every Tuesday at 15:00 UTC (10:00 EST) beginning 30 January. - ATG2 passed the following resolution for input to the next FMG meeting [this is the wording I recorded; the final submission may differ slightly]: - [ATG2 is] Concerned with the isolation of the UNeDocs project in relationship to the ODP and the Forum workflow. - Especially the fact that UNeDocs publishes XML (schemas and code lists) that has not been processed by ATG2 and does not conform to the XML NDR. - FMG should direct TBG2 to ensure that UNeDocs conforms to the workflow procedure and agreed technical specifications of UN/CEFACT.