MINUTES OF MONTREAL UBL TC MEETING -- WORK SESSIONS (See also the opening and closing plenary reports) ================================================================== MONDAY AFTERNOON 17 AUGUST 2009 ATTENDANCE Jon Bosak (chair) Mavis Cournane Mike Grimley G. Ken Holman Zarella Rendon Tony Coates (by phone) Oriol Bausą (by phone) SCHEMA FRAGMENTS We note that the schema pieces prepared by Oriol validate fine; GKH now massaging them to meet strategy doc details. CODE LIST DISCUSSION Each custodian of a code list will be publishing "snapshots" of the list. AGREED that this is what we'll have in each release: the current snapshots. So the gc directory in UBL 2.0 has the snapshots for 2006, the gc directory in UBL 2.0 will have those for 2009, and so on. See Section 1.3.2.2.2 of the Customization Guidelines for a list of the code lists included in 2.0 The default xsl file for 2.1 (created under control of the CVA file) will therefore specify the union of the 2.0-era snapshots and the 2.1-era snapshots. For example, the valid currency codes will be the union of the valid 2.0 codes and the valid 2.1 codes. Our assumption regarding instance level metadata is that the users know what they are doing if they specify a particular version of a code list. For example, the code LEW has meant two different things over the history of currency codes; this is why a user might want to specify a particular version of the list. In the absence of such a specification, they will get the union set. The effect of making a custom default xsl file from a custom cva (including, for example, just the codes from the 2009 version of the code list) would be to screen out instances with obsolete codes. See section 3.1 of the strategy document for an example of this: http://lists.oasis-open.org/archives/ubl/200907/msg00011.html [The strategy document has since been replaced by a revised version: http://lists.oasis-open.org/archives/ubl/200908/msg00020.html ] AGREED that we do not want to publish the 2.0 versions of code lists in the 2.1 distribution; rather, we will direct the user to install 2.x deliverables before installing 2.y deliverables in a sibling directory. This is basically the layout on docs.oasis-open/ubl. We need to explain in the documentation that we assume that users have followed instructions about installing 2.0 and then the 2.0 update (so the 2.0 update is assumed to have overwritten the 2.0 directory); the locationURI in the updated 2.0 codelist files are irrelevant to CVA generation of the xsl file. Users must be made aware that the default behavior is the union set, and that if they want different behavior, they will have to made suitable arrangements (e.g., by using a modified CVA file). If users are not using xsl validation for code lists, then they must be aware that they might need to obtain the other code lists (not just the current snapshots we provide). ACTION: JB to try to write up such an explanation in a draft of the 2.1 documentation for review by the group. ================================================================== TUESDAY MORNING 18 AUGUST 2009 ATTENDANCE Jon Bosak (chair) Mavis Cournane Mike Grimley G. Ken Holman Zarella Rendon AGREED that all file names will be changed to 2.1 from 2.0 and their version attributes updated accordingly, even if their contents have otherwise not changed from 2.0. AGREED that the UBL example files in the "xml" directory are insufficient as a set and that we should obtain examples from OIOUBL or PEPPOL to fill out the set. AGREED to eliminate the former cl/xsdcl directory. ================================================================== TUESDAY AFTERNOON 19 AUGUST 2009 ATTENDANCE Jon Bosak (chair) Mavis Cournane Mike Grimley G. Ken Holman Andy Schoka (by phone) UBL 2.1 DOCUMENTION IN PROGRESS AGREED to retain the qdt file (which happens to be empty but can be used to import enumerated code lists). AGREED to say right after the CVA reference: [CCTS] permits the definition of Qualified Datatypes as derivations from CCTS-specified Unqualified Datatypes. In UBL 2.1, all data type qualifications are expressed in the CVA file [reference]. The [qdt file] in the UBL 2.1 release is included among the schema modules imported by the Common Library and all document-level schema fragments to provide a target for qualified data types in UBL environments using different approaches to data type definition. Figure for UBL Spreadsheet Architecture: AGREED to eliminate qualified datatypes box and arrows. AGREED to add a box to the side (on the right) labeled "Data Type Qualification" with a file box inside labeled "UBL OASIS CVA file" with an arrow down to "Genericode Files." And an arrow to it from both "Document Spreadsheet Model" and "Common Library." (To put it another way: pull the qualified datatypes box out to the right and eliminate arrow to "core component types.") Figure for Spreadsheet Realization: AGREED to reflect the schema structure in 2.1, add to the diagram a single box under the udt box with a single import arrow from the udt box, label the new box "Derived UN/CEFACT Unqualified Datatypes". AGREED to change the figure for UBL Schema Dependencies exactly the same way. REVIEW OF SCHEMA GENERATION STRATEGY DOCUMENT http://lists.oasis-open.org/archives/ubl/200907/msg00011.html Instructions to GKH: - Change language describing the udt module in the new derived udt module. - Change figure 1 to reflect new qdt fragment strategy - (Several other changes to reflect agreement noted above) - Keep gc directory even though cl now has just a single subdirectory ================================================================== WEDNESDAY AFTERNOON 19 AUGUST 2009 ATTENDANCE Jon Bosak (chair) Mavis Cournane Mike Grimley G. Ken Holman PROPOSED NDR 2.1 RULES AGREED as follows: Keep CDL 1-3. No rules need to be added to capture the fact that genericode code lists will be managed using CVA, but add note in the NDR 2.1 doc to mention CVA. CDT5: Remove (no longer applies). A new rule CTD26 will (1) remove the reference to UBL qualified data types and (2) change the UN/CEFACT udt to UBL udt. Remove CTD6, 20, 21, 22, 23, and 24 because we are no longer defining any types in our qdt schema (we used to use CCTS code list types). ELD6: Remove and replace with new rule ELD15 saying that every BBIE shall be based on code.type and shall use the unqualified data type for code. All the IND rules to come out of the 2.1 NDR and its checklist. Section 8 ("Instance Documents") of the current 2.0 NDR should not be included in the 2.1 NDR. All IND to go into the UBL 2.1 documentation as normative. The description of code list validation to be kept "informative." NMS17: Remove and add new rule (probably NMS19) that changes prefix from "udt" to "ccts-udt" and insert "derived" before "CCTS Unqualified"; add explanation in the NDR 2.1 doc. Add new rule (probably NMS20): same as old 17 but replace CCTS with UBL (keeping the prefix "udt"). SSM18: Leave as is (it does define all the UBL QualifiedDatatypes). SSM20: Remove and add new rule (probably SSM22) that replaces CCTS with UBL. No new CVA section in the NDR; this will be in an informative annex in the 2.1 hub doc explaining the contents of the cva directory. COMMENT DISPOSITION: CUSTOMIZATION GUIDELINES PUBLIC REVIEW 2 The only relevant comment received in the public review period was from GKH: A customization should document its calculation model, as Denmark and BII do. An OASIS TC is developing a formalism (test asstertions) that could be used for specifiying calculation models. The question is: do we bring this to the attention of customizers, that they should document the calculation model? http://lists.oasis-open.org/archives/ubl-comment/200905/msg00000.html http://lists.oasis-open.org/archives/ubl-comment/200906/msg00001.html http://lists.oasis-open.org/archives/ubl-comment/200906/msg00002.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00000.html AGREED to add a short new section 1.5 (page 14) as follows: Calculation models The UBL Technical Committee does not prescribe a calculation model that governs how values in instances are calculated (for example, the inclusion of allowances in a line extension amount). Any actual implementation of UBL should document its calculation model via a prose description or a formal description using, for example, the methodology being developed by the OASIS Test Assertion Guidelines Technical Committee. This disposes of the 22 May 2009 comment from G. Ken Holman. AGREED to resubmit the edited CG for another 15-review. Regarding these comments from Stephen Green: http://lists.oasis-open.org/archives/ubl-comment/200906/msg00001.html http://lists.oasis-open.org/archives/ubl-comment/200906/msg00002.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00000.html AGREED not to prescribe a calculation model for 2.1. Regarding these comments from Stephen Green: http://lists.oasis-open.org/archives/ubl-comment/200906/msg00000.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00001.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00002.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00003.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00004.html http://lists.oasis-open.org/archives/ubl-comment/200907/msg00005.html AGREED that these are UBL 2.1 issues for the PSC. Document name: AGREED to change the title to "UBL 2 Guidelines for Customization, First Edition" because: - This edition doesn't apply to UBL 1.0 - This edition does apply to UBL 2.0 - Second Edition will apply to 2.1, but may also apply to 2.2, 2.3, etc. The revised document, ready for its third public review, has been uploaded to: http://www.oasis-open.org/committees/download.php/33852/UBL-Customization100prd03.doc This will be on the agenda of UBL TC meetings in September.