13 Jan 2011


See also: IRC log


Jeff Waters and Don McGarry
Jeff Waters


<scribe> Scribe: Jeff Waters

<scribe> ScribeNick: jeffw

Hi, Piotr

<piotr_nowara> Hi

<piotr_nowara> I'm gonna call in in a sec

Hi, Aaron

<piotr_nowara> http://question-modeling.googlecode.com/svn/trunk/question.owl

<piotr_nowara> http://question-modeling.googlecode.com/svn/trunk/question_example.owl

jeffw: (summary of our agenda, to review Piotr's question format and to review what patterns are needed)
... Piotr, all yours

piotr: I just tried to build something, a prototype still evolving, and I hope we could discuss it, the first link is to pattern itself and the second to the example.
... Inspiration for this pattern, assumption is that question should not be represented only in terms of logic, but question as indiciation of a subject and a desire to inquire some information, borrowed idea from 19th century mathematician

piotr_nowara: Pattern should indicate a question in a context, there are two types of questions in general, first is the question that indicates a subject and property that needs to be answered, like What's your building?
... The second type of question is the question that requires negation of a proposition in question, which requires a yes or no, and it is tricky to represent in an ontology cause it assigns truth or falsity to a proposition
... and the second type of question is difficult to represent because the state of modeling of this type of question cause in context of web ontology language we would need to represent a triple to assign a truth or falsity to it
... I found a work around and I think it is usable. I can provide a link that contains a more difficult, sophisticated example. That example still requires some work. But I can provide a link that lets you look at it.
... There may be other ways of expressing context other than web ontology language that could be used.
... So the lessons learned from modeling question: (1) some kinds of questions can be represented significantly easier than others; So the question_example that I've provide is easy to understand, but the second kind is more demanding;
... (2) You need a domain ontology to model certain types of questions, and (3) when modeling questions you need to deal with static concepts, for example we can have a doctor needing to solve a problem, what drug should I prescribe, normally you wouldn't have such fuzzy concepts in your ontology, but for questions you need them.
... When you analyze those concepts, you discover a more broad context for your domain, which is important. When I modled some questions, it's a kind of disadvantage, but when I thought about it more, I came to conclusion that is is an advantage, it helps you understand your domain better.
... You could design a hierarchy of questions, and another way is to describe certain properties about subject being asked about. The latter good when small ontology and you don't need a catalog of question types.
... One could consider a question parser, having a database schema one could determine whether a question could be answered with data from a given database. That would be a use case. But I would need a good mapping from database schema and ontology domain.
... Last question I want to discuss is some kind of issue within our decision incubator, the question with a link requires some significant domain ontology, without that a proper modeling cannot be done.
... When do we need to introduce this to our decision format? That's a question for us to consider.

jeffw: (jeff summarized thoughts on question text v. triple, and meta-data issue, and need to provide practical format as well as identify challenging issues that don't stop us.)

Summary of Question Prototype by Piotr (which we just covered)

Review of Remaining Needed Patterns

jeffw: We did early on a state pattern, and we also did other early work.
... For this purpose, I will include some links that Eva provided and her summary. Here are the links:

[1] http://ontologydesignpatterns.org/wiki/Submissions:CriterionSetter [2] http://question-modeling.googlecode.com/svn/trunk/question.owl [3] http://ontologydesignpatterns.org/wiki/Submissions:List [4] http://www.w3.org/2005/Incubator/decision/wiki/Use_Case_OLD_Core_Decision_Model#Draft_Solution [5] http://www.w3.org/2005/Incubator/decision/wiki/Final_Report_Decision_Components_And_Patterns

From_eva_email: I think that Piotr has some really nice patterns that he developed [1-2], and we have the list-pattern [3] that came out of WOP2010, and we have the draft models of decisions from this summer that you, Jeff, and me were working on [4]. In addition we have your "core decision patterns" [5] that we could review again.
... I think that what is essential is to produce the following things: 1) a minimal decision format that contains the minimal set of components of a decision, i.e. our core model, 2) a few extensions of this that shows how it connects to the other patterns and possibly other formats as well, for different use cases.

jeffw: Eva also suggested that there are tools for reengineering XML into RDF, and those tools to the XML format.
... Of the core components on the wiki, http://www.w3.org/2005/Incubator/decision/wiki/Final_Report_Decision_Components_And_Patterns
... I will suggest that we take a look at these and that we shoot for seeing if we can't a representation of each of these formats, but I think we need a owl representation of each of these components.
... I think if we each think about that and then provide some submissions of example of complete decisions represented using those draft components, that would be good. Does that make sense?

Aaron: Sure, I think a complete representation of a decision would be good.

piotr_nowara: I also have a complete representation that I haven't yet finished, but I would be happy to send an owl file and you can take a look before the next the meeting.

marion: I sent you, Jeff, several calls for papers and I'm learning and listening.

jeffw: So that takes us to the end of today's meeting, in conclusion I see three initiatives for us as we begin to wrap up our incubator over the next three months:
... (1) A draft OWL-based representation of a complete decision format, or at least the core components of the decision format with some example extensions;
... (2) A paper summarizing our work (the format, the lessons learned, the use cases, any examples of its use) which can be a follow-on to our position paper (http://ontologydesignpatterns.org/wiki/images/0/0f/Pap3.pdf)

(3) Any example of instrumented tools or tool designs that utilize the format to show its practical application.

jeffw: These are stretch goals, but good goals for us because they give us practical substantive products which we can utilize in our ongoing work individually or collaboratively.
... For next time, I'd like to suggest that we each take a stab at OWL format and examples that illustrate a complete decision representation, and then present and discuss at our next meeting.
... Piotr, thanks for you submission of your Question pattern and your overview and discussion of lessons learned.

<piotr_nowara> I'll try to complete my example I started to worked on

jeffw: Aaron, it would be great to include any security components that you think would make sense, so let me know if you have any suggestions. I understand that you are clarifying what you can share and then we can proceed along those lines. Thanks.

<piotr_nowara> It should be ready untill the next meeting

<piotr_nowara> it inludes the question and criterion pattern

jeffw: Sounds good, Piotr, and I will provide an RDF/OWL representation of the XML format but also trying to reuse our patterns and generate any new patterns.
... I guess that's it for today. Thanks to all for attending and I look forward to our next meeting!

<piotr_nowara> Thanks, Bye!

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2011/01/13 16:14:47 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found Scribe: Jeff Waters
Found ScribeNick: jeffw

WARNING: No "Present: ... " found!
Possibly Present: Aaron From_eva_email ScribeNick jeffw marion piotr piotr_nowara
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Agenda: http://www.w3.org/2005/Incubator/decision/wiki/Decision_Mtg_20_Agenda
Got date from IRC log name: 13 Jan 2011
Guessing minutes URL: http://www.w3.org/2011/01/13-decision-xg-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]