Minutes for 20 February 2008 URW3 Meeting

Attendees:

• Paulo Costa
• Nicola Fanizzi
• Francis Fung
• Mitch Kokar
• Kathryn Laskey
• Ken Laskey
• Thomas Lukasiewicz
• Trevor Martin
• Michael Pool
• GiorgosStoilos
• Peter Vojtáš

Agenda:

1. Approve minutes from previous meeting. Note, the minutes are linked from http://www.w3.org/2005/Incubator/urw3/#Minutes
2. Review status of action items from previous meeting
3. Editing team report on status of final report and completion status of work assignments
4. Discussion of conclusions and recommendations
5. Discuss path to completion of final report, including wiki to html conversion, issues tracking, and finalizing the report
6. Other business

Scribe:  Kathryn Laskey

Discussion summary:

1. Approve minutes from previous meeting. 

   Minutes from previous meeting approved.

2. Review status of action items from previous meeting
   
   Previous Action 1: Ken to investigate how to use Tracker and share with rest of us.
   
   Done. A URW3 Tracker instance has been set up. Ken sent out an email explaining how to use Tracker.  
   
   Previous Action 2: Kathy, Peter and Mitch to meet by phone to do speaker use case and go over uncertainty ontology.
   
   Done. Meeting occurred, we discussed ontology, and made some minor changes. We agreed that uncertainty annotations support automated processing of web resources.  Peter attached a refined ontology to a use case.
   
   Previous Action 3:  Mitch to continue working on the speaker use case.
   
   Done. Speaker use case is essentially complete, but did not attach UncAnn tags.  Action 1: Mitch to add UncAnn tags to speaker use case.
   
   Previous Action 4:  Thomas will follow up with Vipul and get firm date to flesh out medical use case.
   
   Done. Vipul has filled in use case.
   
   Previous Action 5:   Ernesto and other contributors to hold recommendation and appointment making have a conference call.
   
   Still trying to hold telecon. Action item carried over as Action 2. 
   
   Previous Action 6:  Editors bug authors of use cases to annotate them.
   
   The editors have been bugging people; will get to status later.
   
   Previous Action 7:   Editing team have telecon to decide recommendation for content and presentation of section 5.
   
   Suggestion was make to include links to the wiki in the final report rather than putting them all into the html document. Some use cases are quite complete in the wiki page, but not in the final report. Ken asked for specific instructions on what to do with use cases. Action 3: Editing team to send instructions with a link to a good use case with annotations, and say to please make yours look like this.  
   
   Previous Action 8: Mitch to set up wiki pages for benefits and recommendations and send emails requesting discussion.
   
   Done. Mitch started a wiki page for benefits and recommendations. No one has contributed.
   
   Previous Action 9:  Ken will page-scrape today's wiki report and save it.
   
   Done. Ken printed it to a pdf file and sent to editing team.

3. Editing team report on status of final report and completion status of work assignments

   • Final report wiki draft is at http://www.w3.org/2005/Incubator/urw3/wiki/DraftFinalReport.  
   • Sections 1, 2 and 3 are ready.
   • Section 4 has a link to a book chapter Paulo, Kathy and Thomas wrote. It is visible only to the members because of copyright restrictions.  Action 4: Paulo seek copyright permission to include excerpts from book chapter.
   • Section 5 is the use case descriptions; people have been working on this. Review of use cases is on today's agenda, with lower priority than Sections 6 and 7.   
   • Sections 6 and 7 are on the agenda for today's discussion. Section 6 is Trevor's contribution. He would like others to add to it.
   • Action 5: Ken to remove write permissions from wiki pages after March 2.
   • Nico noted that the dates were incorrect on the assignment page. 
   • Paulo has a table of all the use cases and their status. 
   • Mike Pool suggested that the use cases should have a field on standardization: what needs to be standardized, and how standardization would help the use case. It was noted that this is a good idea, but the suggestion has come too late in the process. in Peter noted that in his 5.3 there is recommend those aspects that are considered most important to be included in a standard representation of vagueness and uncertainty.  
   • Action 6: Paulo on behalf of editing team to follow up coming up with section 5 format by sending format to all use case authors, with email pointing to specific weak points or omissions in use case that needs to be addressed
4. Discussion of conclusions and recommendations.
   
   
   • There was discussion of what our conclusions and recommendations should be regarding standardization. Mike suggested each author notes what they see as use of standardization re their use case.  Ken said he thought standardization conclusions are more cross-use-cases rather htan within individual use cases; Nico agreed. There was discussion about benefits of standardization, and whether standardization was needed. Mike thought the case for standardization was not clear.  Mike thought that web applications doing reasoning with uncertainty doesn't require standardization; others noted that communicating uncertainty information among web services could benefit from standardization.  
   • Mike gave example of requesting hotel reservation; system comes back with ranked list. System might do uncertainty reasoning, but there is no need for uncertainty representation except inside the tool doing the reasoning. Therefore, why are standards needed?
   • Nico suggested the user may want (be able) to interpret the scores that produced those ranks.
   • Perhaps the information is so varied that coming up with standards is a lost cause.  XGs are supposed to think about areas where you might want standards, but sometimes you realize in process of standardizing that you don't understand problem well enough yet, and standardization is premature.
   • Kathy said she thought there are cases where internal uncertainty processing is enough, but others where maybe it would be useful to standardize. Example of appointment making: want to know how long it will take to get from work to the appointment. Maybe there is a trraffic forecasting system that the appointment making system calls to estimate time it will take to get there. In this case, the traffic forecasting system might output a probability distibution on travel time, and the appointment system will use it. So will a system for suggesting a route from A to B, and so will other systems. Therefore, standards for probabilities are needed.
   • Ken said there are 2 portions to the standards question. One is the rationale for standardizing; the other is whether it is possible. Need to ask what is represented; whether the standards are accessible, practical, useful. We should go back to the use cases and think about what standards are needed.  Regarding need to standardize, Mitch said we can argue that that w/o standards exchange would not be possible, and standard doesn't already exist; state that what exists now isn't sufficient.
   • Trevor says what's listed in Section 3 would be basis for standardization.
   • It was noted that there is a need to include not just probabilities, but also  information about who made the assessment and context in which assessment is made. Nico noted that's true of most web reasoning, it's nothing unique to uncertainty reasoning. 
5. Discuss path to completion of final report, including wiki to html conversion, issues tracking, and finalizing the report
   
   
   • Everyone read Sections 1-4 and make suggestions.
   • Trevor will update section 6. Action 7: Trevor to go through IRC log and make first stab at pulling together a coherent view of discussion
   • Draft of Section 7 is at http://www.w3.org/2005/Incubator/urw3/wiki/Recommendations
   • Editing team will send template for for Appendix to everyone.
   • Everyone should move their use cases from wiki to final report appendix. 
   • Section 5 has only a brief description of each use case. All details will be in Appendix. 
   • Paulo will convert final report to html after wiki closes and before our next meeting. Everyone read it before the meeting.
6. Other business

None.

Meeting was adjourned.

Action items

1. ACTION (1): Mitch to add UncAnn tag to speaker use case.
2. ACTION (2):  Kathy to set up appointment and recommendation use cases telecon with authors of those use cases.
3. ACTION (3): Paulo and the rest of the editing team to point to specific example of how use case should look for section 5. 
4. ACTION (4): Paulo seek copyright permission to include excerpts from book chapter.
5. ACTION (5): Ken to remove write permissions from wiki pages after March 2.
6. ACTION (6): Paulo on behalf of editing team to follow up coming up with section 5 format by sending format to all use case authors, with email pointing to specific weak points or omissions in use case that needs to be addressed.
7. ACTION (7): Trevor to go through IRC log and make first stab at pulling together a coherent view of discussion