UBL-Ontology Project Kick-off Conference Call - Wed 2003-03-12    (5SN)

Agenda    (5TR)

  1) Welcome    (5TS)
  2) Appointment of secretary to take minutes (use Wiki?)    (5TT)
  3) Roll-call of participants (w/ VNC?)    (5TU)
  4) Overview of UBL Ontology Project    (5TV)
  5) Participant introductions    (5TW)
  6) Communication logistics & work protocols    (5TX)
     a. VNC / screen sharing    (5TY)
     b. Wiki    (5TZ)
     c. Threads / subject lines / use of Purple Numbers    (5U0)
     d. Posting of HTML messages / checking posts    (5U1)
     e. Upload to file-sharing workspace (ftp)    (5U2)
     f. Regular conference calls    (5U3)
  7) Other business    (5U4)

Proceedings    (5U5)

Meeting Minutes for UBL Ontology Project Kickoff 3/12/2003    (5U6)

1. Participants: <see above />    (5U7)

2. Kurt gave an overview of the concept for the project.    (5U8)

- Began with a history of UBL.    (5U9)

- Business Information Entities (BIE) are standard abstract classes.    (5UA)

- Attempting to build and execute XML Schemas around the core BIEs.    (5UB)

- The Ontolog forum started as an adjunct to the UBL effort.    (5UC)

- What are the benefits of formalizing the UBL concepts in an Ontology?    (5UD)

- Important question: can we do this project in a distributed and volunteer manner?    (5UE)

3. Patrick Cassidy commented that the only progress the IEEE-SUO Standard Upper Ontology study group made on ontology development was through a funded effort.    (5UF)

4. Norma Slattery asked the name of the Ontology. Answer: SUMO (="Suggested Upper Merged Ontology")    (5UG)

5. Patrick ported SUMO into protégé. The files are available in directory ftp://micra.com/ontolog/    (5UH)

6. Kurt continued the introduction - - Ontolog has brought together a diverse group of talented people. - Technical discussion has been strong. - We need to refine the project goals and make that a first priority.    (5UI)

7. Patrick added that we need to find a non-trivial application to demonstrate the utility of a UBL ontology, and to permit comparison of alternate possible ontologies.    (5UJ)

8. Peter Yim commented that in OASIS for any spec to get adoption requires three use cases for adoption, of which Boeing is offering one.    (5UK)

9. Patrick asked if the use-cases are publicly available. Answer: Yes.    (5UL)

10. Mike Uschold asked if the group knew of any of the people in Boeing who are working on the use cases. Peter answered that it is Sally Chan who is working on the ebXML CCSD (Core Component Supplemental Document). Sally is a member of the ontolog community, and has actually confirmed her intention to participate in this (UBL-Ontology) project too.    (5UM)

11. Moving on to Agenda Item #5 ? Introduction of participants.    (5UN)

- Mike Daconta introduced himself. Co-authored a book with Leo Obrst. Chief Architect of DOD Virtual Knowledge Base project.    (5UO)

- Peter Yim, with Leo Obrst and Kurt Conrad, convened this forum. He worked with Leo at Vertical Net. Cautious of grand plan for Ontologies; however, believes that UBL is adopting the 80/20 rule to narrow down the domain enough. He works with Kurt on UBL. He will be happy to provide the collaborative tools, infrastructure and processes for this project.    (5UP)

- Norma Slattery: Graduate student at George Mason University. Working on a research area dissertation. Looking at this from a Systems Engineering perspective for Ontology development. Works at Mitre. Much hand-waving and ambiguity when it comes to work in this area. Goal of her dissertation is to identify those metrics and guidelines that would support a decision maker. Kurt asked if she had any guidance for the group? Answer: in collection mode.    (5UQ)

- Patrick Cassidy: Started as Chemist. Retired. Need a good semantic dictionary for natural language understanding. Working to create a standard upper ontology to create a standard catalog of fundamental concepts. Worked at Vertical Net with Leo Obrst. Currently on his own.    (5UR)

- Bob Smith: retired Business professor. Interest in expert systems and AI tools. Willing to put time, money and effort into something that is deliverable to achieve UBL success. ? Peter Yim asked for people to email him with their time zone on the next response to meeting time requests.    (5US)

- Bo Newman. Different approach to ontological studies ? at the human level and organization level ? interested in codifying ontologies. How ontologies migrate and change over time. How do we perpetuate knowledge of multi-millenial time spans. Over 20,000 years. How do we capture ontologies when artifacts are separated by thousands of years. Very interested in the process and how are we going to do this as a diverse, volunteer and unfunded team.    (5UT)

- Kurt : Out of SGML and XML community. Principal question is how do we formalize a language? Helped launch an ontology project in CISCO a few years ago.    (5UU)

12. VNC screen sharing technology. Kurt suggested we set up times when people can get familiar with the technology. Peter agreed that times are acceptable but that an always-on capability is not practical. Peter suggested we set up a timeslot.    (5UV)

13. Wiki use. Peter Y explained Wiki. The Wiki URL is at the bottom of every email now.    (5UW)

- Question asked about the purple numbers. It was noticed that a print out of the content caused many pages to be printed for each content break when printing from the Netscape/Mozilla browser.    (5UX)

- Bo Newman suggested that Wiki becomes an ontological development tool in itself by creating new pages for each new concept.    (5UY)

- Peter Yim instructed the group on the use of Wiki and what a ?WikiWord? is. Every time an upper-camel-case word is introduced it is supposed to link to a separate page that describes it. Peter walked through some Wiki functionality.    (5UZ)

- A question mark is generated for every WikiWord that is not linked to an existing page.    (5V0)

- You can force a page (without a wiki word) by using square brackets around a url and then the word/phrase that is the new page.    (5V1)

- Bo Newman added that all the pages are referenceable and the ability to find out what is referenceable within a page. Asked if Peter could talk about that.    (5V2)

- Peter asked if there are any questions on WikiWord? No questions.    (5V3)

- We are using ?purple wiki?. This is built on UseModWiki (the same flavor of wiki that the original WikiPedia was built on.) The key concepts are based on the power of a read-and-write web, the ?wiki word? concept (automated link creation), and purple numbers (automatic creation of node references for each paragraph or lits item within a page.)    (5V4)

- Purple numbers ? after each paragraph the system will automatically put an "nid" (node-id) and an anchor to the top of that paragraph. The way this will be used ? in essence this allows paragraph level linking. Peter then told how to get the URL of a paragraph to refer to it. Question: is it possible to manually adjust the purple number? Answer: yes. Therefore several related paragraphs can be given the same number. ? Bo Newman stated that the question of context in ontological computing is crucial. We can learn from Wiki in terms of contextualization. We need to leverage the Wiki in terms of how it captures artifacts. Kurt added that it seems that productivity falls off on the email list when the conversation wanders and this is a good use for Wiki. Peter offered a MichaelSchrage metaphor of the wiki as the restaurant napkin. Wiki is also the knowledge base for group information. Shared platform for collaborative editing and project display.    (5V5)

Question: on the ability to generate a list of changes and send those changes to the email list server so people know that changes have occurred on a daily basis. Answer: the RecentChanges link gives you the ability to generate a change page.    (5V6)

Mike D asked if the recent changes can be posted to the list serv. Answer: possible but it is a choice whether people want the increased traffic.    (5V7)

 follow-up action done. (see post -ppy)    (5V8)

14. Kurt asked the time zone of everyone on the call.    (5V9)

15. Peter continued demonstrating Wiki. Focused on the UblOntology project page.    (5VA)

16. Mike D suggested we discuss Project goals for the last few minutes.    (5VB)

17. Before going into that Peter suggests we make future calls short (no more than 60 minutes, hopefully even 30 minutes). Use the forum and wiki to transact normal business. Use the calls to build rapport, sync up, distribute and assign or arbitrate virtual ownership of tasks, and use it to deal with emergencies.    (5VC)

18. Mike Daconta suggested that the BIE class definitions exist, so we need a specific use case for the Ontology to drive our development.    (5VD)

19. Kurt asked if the UBL uses cases could fit this role.    (5VE)

20. Kurt took an action item to get the UBL use cases from the UBL leadership.    (5VF)

21. Mike D offered a potential ontology use case: automated Order/Invoice validation so that two software agents could participate in a UBL transaction.    (5VG)

22. Bo Newman added that the ontology is to formalize and codify the UBL concepts.    (5VH)

23. Peter Y stated that one use of the UBL ontology is to contribute to the context methodology work in UBL.    (5VI)

24. Peter Y also suggested that an UBL Ontology could provide normalization support, as per JohnYunker (see http://ontolog.cim3.net/forum//ontolog-forum/2003-03/msg00006.html#nid03), or support in UN/CEFACT standards harmonization efforts which some members of the ontolog community (SueProbert and MichaelAdcock from the UK) are leading.    (5VJ)

25. Kurt added to Peter Y?s concept that when the UBL schemas are extended you will need to have a contextual framework and formal ontology to be able to successful use the UBL transactions in specific contexts.    (5VK)

26. Mike D summarized that we have two axes of development: a specific application (like software agents) and a contextual framework.    (5VL)

27. Patrick Cassidy agrees that a general ontology needs to be useable in specific domains.    (5VM)

28. Kurt Conrad said that he would like to hear Peter?s comments on the concept of generalization. The issue of ? is it too context-specific or can it be codified in a general fashion?    (5VN)

29. Mike D discussed the tension between bottom-up modeling and top-down modeling.    (5VO)

30. ??? stated that the ebXML approach is taking a bottom-up approach in specific transactions.    (5VP)

31. Discussion occurred on product identification in UBL. Consensus: they can mention a standard code (e.g. a NAICS code number) but it is not mandatory in UBL.    (5VQ)

32. Kurt Conrad wrapped up the call. Good progress on rationale of the project. He asked each person to recall their conversations and post them to the listserv.    (5VR)

33. Peter Y added that he will support the infrastructure on a pro bono basis for now, but he is looking for sponsors. Looking to help Peter pay for the cost of the infrastructure. Pat mentioned that Intel does provide some funding for research and asked if anyone had contacts.    (5VS)

 --Minutes by MichaelDaconta/3.12.03
 --updated by PeterYim/3.12.03-14:58PST    (5VT)