UBL-Ontology Project Conference Call - Thu 2003-03-27 (5ZD)
- Conference Call Details: (5ZE)
- Subject: [ubl-ontology] project conference call (5ZF)
- Comments/Agenda: (5ZG)
- Agenda Comments: ... build rapport; ... sync up; ... distribute, assign or arbitrate virtual ownership of tasks; ... and use it to deal with emergencies. ... Also, seed the dsicussion on how we should get organized to start doing project work. (5ZH)
- Please post you suggested agenda items if you would like to see anything put onto the agenda (5ZI)
- please post or upload any material to be shared to the list, to the wiki or by ftp upload prior to the meeting (5ZJ)
- VNC session will not be available for this session. (5ZK)
- During the call, please browse the wiki page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?Conference_Call_2003-03-27 (5ZL)
- Date:Thursday, March 27, 2003 (5ZM)
- Start Time: 10:30 AM Pacific Std Time (5ZN)
- End Time: 11:25 AM Pacific Std Time (5ZO)
- Dial-in Number: 1-702-851-3330 (Las Vegas, Nevada) (5ZP)
- Participant Access Code: "030327" (5ZQ)
- Attendees: - expecting - (5ZR)
Agenda Ideas (602)
Agenda (603)
Meeting Details (604)
Date: Thursday, March 27, 2003 Start Time: 10:30 am PST / 1:30pm EST End Time: 11:20 PST (reserved to 11:55 a.m. PST) Parties: 20 (605)
Dial-in Number: 1-702-851-3330 (Las Vegas, Nevada) Access Code: 030327 (606)
No VNC will be provided for this meeting. Instead, point browsers to the browsers to the wiki for this meeting: (607)
http://ontolog.cim3.net/cgi-bin/wiki.pl?Conference_Call_2003-03-27 (608)
Agenda (609)
1) Welcome (60A)
2) Appointment of secretary to take minutes (use Wiki?) (60B)
3) Roll-call of participants (60C)
4) ?Wrap up? on goals (anyone who missed last meeting) (60D)
5) Communication logistics & work protocols (60E)
a. VNC / screen sharing b. Wiki c. Threads / subject lines / use of Purple Numbers d. Posting of HTML messages / checking posts e. Regular conference calls (60F)
6) Summaries of Proposed Methodologies (60G)
a. Stanford / Protégé Ontology 101 b. Guarino & Welty OntoClean c. Role of Wiki d. Other approaches (Gomez, et. al.) (60H)
7) Methodology issues (60I)
a. Strengths and weaknesses b. Completeness (degree of formalization, superset / subset) c. Class modeling see http://citeseer.nj.nec.com/parsons00emancipating.html (60J)
8) Ready for vote on methodologies? (60K)
9) Use cases (60L)
a. What used for? b. What needs to be included? c. Assessment of existing UBL use cases (60M)
10) Tasking (60N)
a. Previous assignments b. New assignments (60O)
11) Other Business (60P)
Proceedings (60Q)
1. Participants: (60R)
- KurtConrad (60S)
- AdamPease (60T)
- SamHunting (60U)
- BobSmith (60V)
- BillMcCarthy (60W)
- MonicaMartin (60X)
- MichaelDaconta (60Y)
2. Kurn Conrad reintroduced the question of goals. (60Z)
3. Adam Pease stated that he is interested in Ontology content definition. Specifically logical axioms. (610)
- He asked if anyone has started creating formalisms. (611)
- He mentioned that the current UBL spec just has english definitions. (612)
4. Monica explained the plan/process for the UBL committee on their plans to refine the business information entities. (613)
5. Kurt added that the UBL committee started out with a business vocabulary (existing library of terms). This consists of core components and business information entities. Mostly developed at the level of vocabularies. (614)
6. Adam asked is their a compact specification of the business information entities? (615)
- Monica mentioned that their are business information entities, spreadsheets and uml class diagrams. (616)
- Adam said he is looking for a set of UML diagrams that specifies the constraints on the information models. (617)
- Monica mentioned there are also spreadsheets. (618)
- Kurt thought that Adam would be interested in looking at the Excel spreadsheet. However, the richer semantic properties may not be in there. (619)
7. Adam stated that there is an advantage for this domain specific ontology to be an extension of an upper ontology. So we need to put it on the agenda to pick an upper level ontology. (61A)
- Kurt asked what the candidates for that are. (61B)
- Adam said that the only free upper ontology with formal rules is the Suggested Upper Merged Ontology (SUMO). At http://ontology.technology.com (61C)
- Kurt asked how this is related to CYC (Adam said that is propietary, although the taxonomy -- minus the rules -- has been released publically) and John Sowa's ontology (SUMO did start with that). (61D)
- Kurt asked if their is a seminal paper that describes the design. Answer: yes on the website. (61E)
- MichaelDaconta agreed that this should be tied into an upper ontology. (61F)
8. Kurt asked if anyone else had suggestions for goals. (61G)
- BillMcCarthy suggested we also take a survey of existing domain ontologies for reuse. (61H)
- Adam said that some of that has been done in creating the SUMO. (61I)
- Adam said that they have also been creating a bunch of specific domain ontologies that subclass the SUMO. (61J)
9. Next agenda item: Communication logistics. Addressing problem areas in the infrastructure. (61K)
- No one felt there was any problems. (61L)
10. Next agenda item: Ontology building methodology. (61M)
- Kurt reviewed the current candidates. (61N)
- Adam offered his opinion of the Welty/Guarino ontology methodology (OntoClean). The six guidelines are good guidelines but not as rules. The doctrine of essentiality is a great guideline but if adhered to without exception can cause problems. Thus we must treat this with a heavy dose of common sense. (61O)
- Do we have enough consensus on methodology to move forward? (61P)
11. SamHunting mentioned that common sense is good and that we should not get into religious battles. (61Q)
12. KurtConrad mentioned a paper on citeseer (URL is on wiki) that warns against exclusively relying on class modeling for this. (61R)
- SamHunting seconded this. He felt the paper is superb. (61S)
- SamHunting mentioned that it would be great if the formal ontology languages would be able to map into Wiki or that Wiki could understand. (61T)
- MichaelDaconta believed that may require custom coding. (61U)
13. AdamPease mentioned that there really is no widespread consensus on a specific methodology. There is no standard methodology for ontology development. There is no cookbook procedure and that does not relate to a lack of quality. (61V)
14. KurtConrad feels we need a methodology czar. (61W)
- MichaelDaconta thinks we should just pick one and get started. (61X)
- KurtConrad asked if we ready to settle on a methodology? (61Y)
- AdamPease thinks it is premature. (61Z)
- AdamPease felt that step 6 was too specific (define the facets of slots). (620)
15. LeoObrst joined the call. (621)
- KurtConrad gave him a summary of the discussion. (622)
- Leo felt the Ontology101 approach was ok and probably simpler than methontology. (623)
- Leo added that Ontoclean was fairly esoteric. That should not be confused with methontology. He concluded that a simple ontology that works would be fine. (624)
16. Kurt rephrased the position. We can use the Ontology 101 method up to step 6 and at that point we need to stop and pick a set of tools. (625)
17. Adam is impressed that this group is very functional. (626)
18. The group adopted the following position: (627)
- We use Ontology 101 up to step 6 and then decide on specific tools/languages. (628)
19. Kurt brought up the article again that relying too heavily on class modeling can skew the results. (629)
- Leo mentioned that most of the time you do focus on class modeling first. (62A)
- Kurt agreed we can table that issue for today. (62B)
20. Next agenda item: use cases. (62C)
- LeoObrst stated that the UBL spec covers a lot of ground. We could either use their use cases or prioritize our modeling (62D)
based upon our own priority for use cases (our own goal). One use case may be: a company wants to buy something for another company via a particular type of transaction in a specific context. For example a buy versus a sell. (62E)
- Leo continued that usually the model must be guided by a purpose or use case. Otherwise you may be modeling (62F)
- MichaelDaconta suggested the use case where an agent is able to validate that an Order and Invoice are semantically the same. This is beyond just the order # being identical. (62G)
- Adam mentioned that there are many levels to the validation scenario. (62H)
- A discussion ensued about the suitability and difficulties in implementing the validation scenario. (62I)
- Monica Martin stated that there is some work going on relating to the componentization of products for product identification. To answer the question of identifying products by function instead of by name or product id. People asked if this could be posted to the list server. (62J)
21. Pat Cassidy joined the call. (62K)
22. Adam Pease suggested that defining scope is critical. He related a story of a group that has been working for several years on just an ontology for what appears on the front of an envelope. (62L)
23. Kurt asked if the UBL use cases enough? (62M)
- MichaelDaconta suggested that the UBL use cases are different than the uses cases for the UBL Ontology. (62N)
- Leo noted that they are related. (62O)
24. Meeting was adjourned. (62P)