UBL-Ontology Project meeting on "Modeling / Formalizing Representation of Core Component Types" - Thu 2004-03-11 (1H5)
- Subject: [UblOntology] CctRepresentation project conference call Thu 2004-03-11 (1H7)
- Comments/Agenda: (1H8)
- Date: Thursday, Mar. 11, 2004 (1HG)
- Start Time: 4:00 PM Pacific Standard Time (1HH)
- End Time: 5:55 PM Pacific Standard Time (1HJ)
- Dial-in Number: 1-702-851-3330 (Las Vegas, Nevada) (1HK)
- Participant Access Code: "040311#" (1HL)
Agenda Ideas (1I1)
- to get ready for the meeting, please review the following discussion threads on the [ontolog-forum] that began with: (1I2)
- (please post your suggested agenda items or discussion topics below) (1I7)
- What does it take, to convince people that we need an ontological engineering approach, to properly capture the needed semantics? (1I8)
- Does the question -- in our case here, of (a) representing ebXML Core Cmponent Types, and (b) ontolog's proposal of our UblOntology methodology -- boil down to the two following questions? (1I9)
- (1) "What is the language that the semantics is going to be expressed in?" (1IA)
- (2) "do we need KIF, or can UML adequately/fully represent the semantics?" (1IB)
- If so, let's explore these; if not (or not only), then what else is pertinent? Let's list them and explore them too. (1IC)
Agenda & Proceedings (1ID)
- 1. Opening / Overview / Adopt today's agenda (1IH)
- note of thanks to Tim, for starting the subject and the investigation. (1II)
- 2. What are the issues? (1IJ)
- ebXML CCTS related (1IK)
- code vs identifier (1IL)
- quanitty and measure (1IM)
- possibly smaller core set (1IN)
- UBL related issues and UBL impact (1IO)
- ebXML CCTS people liked the UML diagram (1IP)
- helped correct typos and suggested editorial changes (1IQ)
- recognized that the schemas were oversimplified (in UBL v1.0); have not captured all the CCTs properties (1IR)
- in retrospect, could have gone into a deeper, richer structure (1IS)
- [ontolog] or UblOntology project related (1IT)
- ontolog is in a good position to come up with assessment and recomendations from an ontological engineering perspective (1IU)
- 3. Discussion (1IV)
- Tim: example - code list "ISO 4217 Alpha" (currency code list) (1IW)
- UBL recognizes 3 types of Codes: (1IX)
- (1) (International) Standards; (1IY)
- (2) Stock Codes (industry group or government agency defined); and (1IZ)
- (3) "Placebo" (something agreed upon, but unknow to UBL) (1J0)
- UBL's delineation between "Code" and "Identifier" is (1J1)
- if it is an abbreviation, it is a "code" (e.g. "USD" is a code; "bar code" is not a code) (1J2)
- an "Identifier" is something that uniquely identifies within an aggregate, object or structure (e.g. in a "bar code identifer" or "product identifer") (1J3)
- see the whitepaper from UBL (1J4)
- another issue, for example, is 'why is "Amount" not a "Measure", and "Temperature" is' (1J5)
- Tim: the CCTs could possibly be represented at the core by a smaller set; UBL is providing a proof-of-concept implementation (1J6)
- 3 groups have agreed to share the same CCT schemas: UBL, OAG & UN/CEFACT-ATG-2 (1J7)
- "unspecialized data types" vs "specialized data types" (1J8)
- Peter: suggest that Tim (UBL/LC), Nenad (OAG/NIST), Sue/Monica/John/Bill from UN/CEFACT and some of us at ontolog and try to put together a report of Findings and Recommendations for UN/CEFACT (1J9)
- Adam solicited Tim/UBL to review his mapping and comment to that work (see: http://ontolog.cim3.net/forum/ontolog-forum/2004-02/msg00045.html) (1JA)
- Tim agreed to do it (1JB)
- Nenad: NIST does provide input into the convergence community; seconds the engagement of multiple standards groups to work on this collaboratively (1JC)
- Tim: question - is it a difficult process to go from KIF to UML (1JD)
- Adam: I took a look at OCL (UML's Object Constraint Language); and the answer is "yes". OCL is not ontologically neutral at all (1JE)
- Tim: UN/CEFACT's commitment to UML is in the class diagram, and not to OCL (1JF)
- Adam: to go from KIF to a UML class diagram would be trivial, because it is a lossy translation; we can still represent the additional semantics in KIF and still be compliant (since they are not committed to OCL) (1JG)
- 4. Deliverable(s) (1JH)
- Can we reach some concrete and actionable conclusions (today or in relatively short order) and provide that as feedback and recommendations to UBL and ebXML-CCTS. (1JI)
- Tim: these (CCTs) may possibly be represented by a smaller set; UBL provides a proof-of-concept implementation (1JJ)
- we'll form a small task force to work on: "Findings and Recommendation for ebXML Core Component Types" (need to wordsmith this too) (1JK)
- clarification, rather than redefinition (1JL)
- get people from all concerned parties involved (1JM)
- Process and Timeline: (1JN)
- clarify UN/CEFACT's intent, where we feel it is ambiguous - try approaching Mark Crawford, editor of CCTS, as the point person (1JO)
- comments to Adam's mapping -- one month (1JP)
- we'll try to take one month to draft the deliverable (1JQ)
- and another month to clean it up for presentation (1JR)
- total time: 3 months (1JS)
- try to match that to the UN/CEFACT timeline (e.g. so we can formally present it at one of their face-to-face meetings) (1JT)
- they are now meeting in Bonn this week; the next meeting is probably 6 months away, probably also in Europe. (1JU)
- Alan Stitzer is probably a good contact regarding a submission (1JV)
- Call ended 2004.03.11 6:15pm PST (1JW)
- will continue the discourse on the forum. (1JX)
- Peter to make post about the new task force/project and start a project page on the wiki for it. (1JY)
--
minutes taken by PeterYim in real time
ppy / 2004.03.11 18:21 PST (1K1)
