- Presentation and discussion on contribution
- Requirement discussion
- Further work
1.2 Roll Call
| Anthony Coates
| Jim Wilson
| Scott Hinkelman
| Tom Glover
| Douglas Hill
| Günter Stuhec
| Nada Reinprecht
| Øyvind Aassve
1.3 Contribtion discussion
Tony Coates presented his contribution - see contribution paper.
- Ontology - overused term, but important, focus on relations - needed in some industries like finance. A solution that is enough, but not more complicated than necessary. But the solution can also not be simpler than the problem we are trying to solve.
- GS1 stress importance of backwards compability.
- Simple metamodel, might have to add identification/groups.
- Tony stresses importance of real worls examples in order to check the solution. Here is my business context - how can we describe it. Avoid purely theoretical exercise.
- Important to avoid solutions where you have to name things very well, very difficult to succeed with in the long run.
- Run-time - not fully fleshed outwhat all these opportunities might be. Run-time requires context information in the message - not only in the schema.
- A certain level of inferencing necessary. Not overboard - full knowledge based system. If Texas can infer US without listing up every possible node. Problem with scope of inferecing - can be enourmous - model management issue. Only list the things we need.
1.4 Requirements discussion
Requirements discussion on Context Methodology:
Decision: Requirements deadline postponed until beginning of May on request of IBM. IBM may at this date if not contributing indicate a day of possible contribution that might be later than requirements deadline. Accepted by the group if delivered reasonably quick. GS1 and SAP are also considering contributions/ submissions on requirements.
SAP contributed with the following requirement proposals:
Req.proposal: Context should be part of SBDH in order to convey the context of the message to the partner.
Req. proposal: Should describe in which language catgories/ constraints should be defined.
Ø Comment: where is the boundry of UCM. Model and process context, but not application (do i allow this rule, this field). UCM spec should not treat these things individually - cannot be complete anyway.
Req. proposal: Possibility to define super- and subcategorization structures.
Req. proposal: Technology representation should be provided as part of the project (might be done by ATG)
Questions to be discussed.
- What should be contextualized? Modelling context independent of artifact it is applied to. UCM should not be responsible for content of the metamodel (i.e. categories and values).
- Trigger behaviour - sending document from Germany to England means sending from one context to another. Possible run-time usage?
1.5 Further work
Weekly meetings: Thursday 08:00-09:00 CET and Thursdays 18:00-19:00 CET every other week.
Need to go through proposed requirements and formulate them for the Requirements document.
Go through wording in existing Requirements document. Wording should be more what and less how.