Attendees: Lisa Seaburg, Joe Chuisano, Terry Schager, Sue Probert, Arofan Gregory, Sally Chan, Debbie Quezadaz, Gunther Stuhec.
1. Welcome from Chair and appointment of Secretary to take minutes.
2. Acceptance of previous minutes. - Quorum not established - deferred to next meeting.
3. Items under review - to be completed by this meeting
This item refers to the second draft of the Methodology paper sent by Tim McGrath to the listserv on 30/7/02 (reference archive message http://lists.oasis-open.org/archives/ubl-lcsc/200207/msg00028.html).
Lisa asked if Tim had set a timeframe for comments - there was none. There was no more discussion on this document at this point in the meeting, as we did not have a status on the document.
Sue reported that there was no movement in this item. Tim will include this as a position paper for our next distribution.
Sue elaborated that she has questions as to the differences between our approach and the approach that the CCSD Team is taking. We also need to incorporate the discussion from the last Core Components specification meeting in Washington, D.C. - so there is a dependency on those changes (the changes to the Core Components specification). Sue stated that the agreements made at that meeting were in line with what the CCSD Team is doing in terms of definitions. Arofan added that the changes were more than just definitions.
Sue asked if there should be a conference call on this. Lisa felt that that was a good idea, and mentioned that we would revisit this at the end of the call (if time permitted).
Neither Gunther nor Mike were present at this point in the meeting (Gunther joined later). Item was skipped.
Lisa stated that she received no other comments - therefore, she will put it forward as is, unless someone objects (no one objected). There was also no response from the NDR Team, and we are not sure if that means acceptance or that the Team did not have a chance to review the model yet.
Lisa also mentioned that there needs to be some cleanup on definitions - Mike will be helping with that.
Lisa clarified that the Guide For Reviewers document is intended to inform reviewers of the types of feedback we are looking to them for, what template to use for feedback, and anything that we need to say up front (this is a work in progress, etc.).
It was also mentioned that the Guide For Reviewers and the Methodology document have some overlap - the Team felt that this was approach, given that the two documents are for two different audiences.
This time around, we will have more reviewers because the Liaison Subcommittee now has more members. Thus, we will reach a wider audience.
There was a discussion on timeframes, and the possibility of moving the distribution deadline back to afford us more time to prepare. Jon Bosak has definitively stated that the deadlines are solid - therefore, we cannot move them. Lisa offered an assessment of our current situation from the QA perspective:
4. Work Plan - Review and Update
Gunther's Perl script: Requires update, because it is using the old NDR rules. Gunther will have this updated by end of this week (Sunday, 11/8/02) [ACTION ITEM].
QA Status: Lisa stated that QA has gone through naming and definitions, and some structure (Mike has worked on structure). She asked Sue for a timeframe as to when her portion will be complete - Sue is aiming for end of week for a first cut review of the definitions (Sunday, 11/8/02) [ACTION ITEM]. Lisa asked Sue to let her know when this is complete, so that she can update the workplan table [ACTION ITEM].
In terms of the spreadsheet, Lisa is in the process of going through everything that Mike sent and doing a cut/paste into the spreadsheet of the new UBL definitions that he wrote.
Sue and Lisa discussed whose work Sue's work will build upon - i.e. whose spreadsheet should Sue add her work to (as order is important here). The resolution of this was that Lisa would request Alan's version from him, and send it to Sue [ACTION ITEM] so that Sue can build on top of Alan's version. Sue will add a column to Alan's spreadsheet with a header of Comments from Sue, or something similar.
Lisa is not sure of Monica's status. Since Monica was not on the call, Lisa will contact her [ACTION ITEM].
Arofan asked if we were planning to publish release notes in the distribution. Lisa felt that we should, and thought that the Reviewer's Guide should cover this.
Arofan pointed out that since we are waiting on the CEFACT Transportation group to perform a review of our transport structures, we should state this in the release notes.
Sue to craft a statement to include in the release notes expressing this [ACTION ITEM].
Lisa to contact Tim on all of the outstanding items for the distribution (release notes, etc.) [ACTION ITEM].
Lisa also asked Gunther to ask Frank if he could send LCSC Team an e-mail on his Part Numbers idea [ACTION ITEM].
5. Schedule for next release
More Discussion about the timeframe follows:
Sally has volunteered to create a stylesheet. In order to do so, she needs a sample XML instance document - which is dependent on Gunther's script being completed. Gunther will generate a schema and create/send an XML instance document as soon as possible, to keep Sally moving forward [ACTION ITEM].
Does our metamodel accommodate global attributes?
Do we intend to address the idea of presentation of markup as an attribute? (e.g. specify font size for a particular BIE)
Elements vs. attributes (Gunter's position paper);
Sue inquired about the changes in definitions, and whether the work will build upon the foundation of updated definitions. Lisa responded yes, and pointed out that we can run the script now with our current definitions, and re-run it later with the updated definitions. This will keep us moving forward.
Gunther reiterated his need to update the script with the most recent NDR Team rules. Lisa to contact Eve and ask her to send all necessary information to Gunther [ACTION ITEM].
Timeframe for upcoming distribution:
The QA Team is getting a lot done - the fact that it is a smaller team has helped;
Out of 400 elements, approximately 20 names were changed;
All definitions are cleaned up, and some typos were caught;
However, we still have the same flat design - i.e. structure has not changed from the Order model
This last point sparked a discussion on structure. Arofan mentioned the Containership paper that he wrote (reference archive message http://lists.oasis-open.org/archives/ubl-lcsc/200207/msg00030.html). Lisa stated that since the paper has not yet advanced enough, it is not part of the QA process.
Arofan expressed a concern that if containership was not done now, we will hear the same comment on flatness in the next go-around. He also mentioned that some of the comments from the first review were not yet dealt with - therefore, we need to make that clear to reviewers so that they know that they are still pending.
Lisa mentioned that Tim's Methodology paper does reference the concept of grouping of elements, in section 3.3.7 (on normalization). We began to look at this section.
Gunther referenced his paper on how to generate object-oriented message. He felt that this would be a nice method to use for message design. Arofan asked if that paper could be used in the NDR Team discussions on element reuse - Gunther concurred, stating that a lot of rules are needed for how to make aggregates reusable.
A discussion of Party information and containership then took place:
Arofan discussed the concept of a single Party container element that would group multiple party element groups. With this technique, a processing program or application could simply access the Party container and pick up all party element subgroups as well.
Otherwise, if there were 15 maximum possible types of parties (for example), a processing program or application would need to run a counter and check if it had reached the 15th party element subgroup for each iteration.
Gunther pointed out that this type of container is not necessary, as a single XPath expression could collect all Party elements. Gunther also stated that if you would like to create class diagrams, all Party objects are object classes - however, you could not generate an object class for the container.
Arofan agreed with this from the OO data binding perspective; however, he does not agree with it for 99% of the XML processing that goes on in the world.
Gunther also stated that if we create a container, naming it could be an issue. Arofan stated that we would call such a container OrderParties. As to how to create a class diagram, Arofan referred Gunther to Dave Carlson.
Gunther agreed to send the OO paper to Dave Carlson [ACTION ITEM].
The Team then discussed in section 3.3.7 of the Methodology document more in-depth.
Arofan asked Joe Chiusano to summarize Tim's descriptions of 1st, 2nd, and 3rd normal forms - Joe did this, using Order and Item concepts as examples.
Joe also pointed out that it is entirely possible to have a model in 1st, 2nd, and 3rd normal form, yet still leave some containership potential open - i.e. normalization does not mean that all elements are grouped to the maximum possible extent.
The Team agreed that it would be good if Tim could describe the normalization concepts in the Methodology document in more plain English form, and perhaps provide some examples. This would help those who do not have as in-depth a relational database design background as others. Lisa to contact Tim on this [ACTION ITEM].
Regarding our library, Arofan stated that we should ensure that our library follows the methodology that we say we are using. He offered three points:
Containership vs. what is in the library - we need to highlight the difference for reviewers (i.e. the flatness of our current structure);
The Methodology and NDR Team do not agree on all points - this should be discussed by Tim and Eve;
Part of this involves whether we include things inline, or reference them
Gunther's OO paper needs to be looked at by both the NDR and LCSC Teams
Arofan's Containership paper needs to be looked at
Update to the Methodology referenced above (plain English);
5b. Timeline update
Lisa reiterated that we should strive to accomplish as much as possible this week, so that we can get our reviews done by next week
A discussion took place on facets:
Arofan asked Gunther if he planned to update his Perl script to accommodate facets. Gunther stated that facets are not yet incorporated into the script, and that this is a complicated issue (we need to define rules for using them). Arofan remarked that this should be included in the release notes, as he heard comments from xCBL folks regarding the lack of datatypes in our first release.
Lisa added that the spreadsheet that she is working with did not include facets, and that she could include some simple facets if Gunther gave her the proper information. Arofan stated that at this point, we should not jam stuff in there without a review by the group, as some of the facets are tricky. Lisa agreed.
Regarding timeline:
Next week, Sally and Lisa will be working on XML instance examples and a stylesheet.
All work needs to be ready for publishing in time for the 16/8/02 Liaison Subcommittee meeting.
On our 13/8/02 call, we should be able to say that most items are complete.
Arofan also mentioned Gunther's Date/Time position paper. NDR Team has completed their discussions on this paper, and this information has changed in the Core Components specification - therefore, anything regarding this in our current methodology is out-of-date.
Rather than re-review and update the model for this, it was decided that this would be included in the release notes as well (i.e. will be done in future). Arofan to help Lisa with this [ACTION ITEM].
6. Reports on status of other UBL subcommittees
We do not yet have a Perl script to create the schema from the spreadsheet;
But, Gunther reiterated that he will complete his Perl script updates by end of this week.
Lisa inquired of Gunther about a requirements paper - Gunther stated that he needed to put this on hold, and would take it offline with Lisa. Arofan and Gunter to send requirements to Lisa for this document [ACTION ITEM].
Arofan asked Gunther if he will be presenting at XML 2002 in December. Gunther stated that due to travel restrictions, he will not be able to attend this conference. Arofan mentioned that he is giving a presentation on NDR, and would be happy to present on Tools and Techniques in Gunther's absence. Arofan to contact Gunther offline regarding this [ACTION ITEM].
Arofan reported that everything is coming together pretty well, and that they are making good progress. He also pointed out that the NDR Team does not have the same pressing deadlines that the LCSC Team has.
Arofan also emphasized the need for reviews of everything substantive that comes out of the NDR Team, and the importance of the LCSC Team remaining in sync with the NDR Team.
Lisa heard that there are discussions of merging the two distributions in the future - Arofan confirmed this. This will help ensure that the two teams are aligned to the greatest extent possible.
Arofan reported that the subcommittee has a draft, but it is not yet released as they are in the process of getting the new membership straightened. He will talk with Lisa about updating the LCSC Team web page for this [ACTION ITEM].
Sue referenced an e-mail sent by Jon Bosak on 13/7/02, and will send it to Arofan [ACTION ITEM].
Mike's assembly document was referenced - Arofan clarified that this document was regarding assembly in general, not context assembly.
Arofan reported that the subcommittee has not been making a lot of progress as of late. He also listed the current members.
Matt Gertner has been away from the subcommittee due to business commitments, and will rejoin soon.
Jon Bosak has been making great progress on this subcommittee, as quite a few members have been added. Lisa listed the new members, at Arofan's request:
Sue reported that EAN-UCC have just announced that they are joining forces with RosettaNet. Arofan stated that SWIFT and EAN are using core components-based models, as are we - therefore, this could be a good opportunity for alignment. Sue agreed, and mentioned that we have James in EAN-UCC on the technical side - but we need to cover the political side.
Sue also expressed that we should work with catalogue groups as well.
Continuing with the EAN-UCC alignment discussion, Arofan asked if there was a shortlist of things that we are doing that we can say, yes, we are doing UCC. He mentioned our external code list technique (dynamic code list access). Arofan proposed that we approach EAN-UCC with the code list definition schema and let them be the first testers - this will help us align with them. We could also capture the idea that by default we will use GLN's
Sue also reported that Jon Bosak is working on an alliance with EAN-UCC, and that she will approach James on this as well [ACTION ITEM].
7. Reports on status of other related projects
Arofan outlined some of the updates to the Core Components specification:
He discussed Tim's CCT/RT comment, and how it was similar to a comment reported by X12. The Core Components Team has updated this to now include a one-to-one correspondence between Core Component Types and Representation Terms. Sue has the updated the cross-reference table she received from Paula Heilig, and will send this to Arofan [ACTION ITEM]. Mike's version is a later version. Arofan asked Gunther to send Mike an e-mail asking for his updated table [ACTION ITEM].
Another major issue was that Core Components did not accommodate object-oriented parent/child capabilities - this was because types could be easily reused. This has been fixed by separating the Property Term from a Basic Core Component (BCC), thus allowing a BCC to have the properties of a parent.
There is also an increased emphasis on use of datatypes - for example, you can declare that a BCC is a 10 character string, rather than simple a string. This approach uses a facets mechanism on datatypes.
Lisa asked if there were any changes that could require changes in our naming methodology - Arofan answered yes. Lisa and Arofan agreed that since these updates cannot be made now (as the updated Core Components specification is not yet released), this will be stated in the release notes.
Gunther asked Sue some questions regarding codelists:
Gunther has found further EDIFACT codelists for identifying organizations (0007 element, for example). He asked Sue which list should be the default for use.
Sue stated that 3055 should be the default, as the others are syntax- bound to an envelope. Sue to respond to Gunter's e-mail on this [ACTION ITEM], as there was no further time to discuss on the call.
CCSD (Monica/Sally)
8. Other Business
9. Next Meeting