24 Jun 2010


See also: IRC log


Jeff Waters, Don McGarry
Jeff Waters


<scribe> Scribe: Jeff Waters

<scribe> ScribeNick: jeffw_

Into and Status

Hi, Eva

<Eva> hi

I'm on the call with my colleague Bruce Plutchak who is joining our efforts

I will send a message to Don to see if he is joining us

<Eva> I am trying to call in...

<don> hey Jeff...my 9am is running a little late

<don> i'll be on the call in 5 min

Hi, Don, no problem

Status I feel very good about our status and accomplishments. We are well-defined in our mission (to represent decisions and decision-making) and our approach (eXtreme Design utilizing ontology design patterns), our available tools (Neon with eXtreme Design plugin and SPARQL plugin), our initial use cases (Information Flow, Situational Awareness, Open Linked Data),

and some key strategic milestones (such as having a decision ontology ready over next 2 to 4 weeks to support upcoming exercises/experiments in late July and August, as well as the milestone of showing the capability of utilizing open linked data for decision-making).

We now have a few key weeks ahead when significant progress on our modeling efforts, if we can achieve it, will be well rewarded at our upcoming events. Also, as we head toward the end of the summer,

we will be half way through our incubator and it will be time to assess our future direction. We will have the opportunit y of considering whether the potential significance of what we are pursuing will help expand the web as is the mission of the W3C and whether representing a decision in the form of a recommendation (standard) by forming a working group is worthy of the effort.

<Eva> sorry, I have some problems with calling in.. join you in a minute I hope

ok, sorry, I was able to call in without problems, Don has also joined us

<Eva> sorry, not sure what the problem is, maybe there is something with the phones here (calling from Sweden at my old university today) because I can connect but it doesnt accept the conference code nor can I get an operator... I keep trying

don: i started editing wiki pages for situational awareness and one of the others
... I put in stuff on command and control and whatever else in there

Modelling Efforts for Particular status

I can begin quickly by noting that I had some significant results to report at our last meeting, which are summarized here http://www.w3.org/2005/Incubator/decision/wiki/Use_Case_Info_Flow_Progress. Essentially, I was able to use the Transition pattern, with some specialization,

for the purpose of modeling decision states including start/stop times to enable tracking of decision state flow and how much time was spent in each state. I created a few classes and properties, then created the instance date for the time intervals, the states, and the transitions, and then I used a SPARQL query to show that I could pull the data from the instances.

So this is my long way of getting around to the point that I didn't do any more on this particular aspect of our modeling efforts over the last two weeks for this meeting. I did do some exploration on metrics with the open linked data, which I can report as our next topic. For now, let's go around.

Eva: I didn't have as much time as I wanted, so I havn't made too much progress on modeling. I thought about the use case and I see two different
... parts of the use case, one is to say we're using this data in this decision and the other is the core decision model to say how the data is used in the decision.
... I had a look at some of the patterns and as far as representing decisions, so we may need to make one.
... I don't see any really clear match among the patterns. Some of the general ones, but if the pattern doesn't add anything, better to make a decision pattern of our own.
... I also read the addition by Jeff, one interesing thing is how you can assess the data using SPARQL and this is something we can think more about, what assessments
... can we make with SPARQL and what can't we do, but on this part not too much progress. I did look for more participants.
... At European Semantic Web, so not too many people interested, but at Open University in UK, I found someone interested
... I sent her some more information and I hope she will join our efforts. And our other option with RPI and Jim's group and I e-mailed more folks.

There is a woman in the group, a PHD student, who is interested in the topic. She could at least follow our work and see. We'll see how this progresses.

Eva: The third guy who Aldo forwarded e-mail to me, he sent something to some W3C list about the decision ontology and I emailed him and told him about our activity and it turns out
... that he doesn't work for a W3C member, but maybe he can get involved
... So outreach activity is my focus recently.

That's great. We should continue the efforts. Any interesting things at the recent semantic conference?

Eva: Main things were quality of data is an issue, most agree that the problem is we need to make applications/reasoners that are robust enough to work
... even with mistakes in the data. I agreed with them. Look at the web, so many pages with text with many broken parts. So our systems need to detect
... what is reliable, what is wrong and use the data anyway. I was expecting to find someone who does something related to the incubator.

I'm so interested in using the open linked data for decision-making that I'm wondering what folks do want to use the open linked data for.

Eva: To put things in context of time or space, plotting things on the map or on the timeline to visualize the data with facts that are linked to the date or position in space.
... This is also about the earthquakes dataset like you were writing about. relating facts to time and space and then visualizing.
... I didn't see any more complex or advanced applications. It's more about showing data to the user.

Don: I will be pursuing the modeling as best I can, but I may have questions.

Eva: I'm here and happy to answer.

Decision Metrics and Open-Linked Data Our next topic is one that was interesting for me. I started thinking about how to utilize the open linked data for decision-making and in particular how it relates to options and metrics.

I summarized my thinking on the wiki with a link on the Use Cases page under the Open Linked Data use case, which points to here: http://www.w3.org/2005/Incubator/decision/wiki/Use_Case_OLD_Explorations_1. (As an aside, I used the abbreviation OLD for open linked data, but it's probably a poor choice since it looks like I'm reporting on "OLD" exploerations when in fact these are "new" explorations.)

I'd just like to walk through what is reported on this page, because I ran into a few issues that I know you all can help me with. First, let's consider an example open linked dataset (earthquakes) and how such a dataset might support decision-making.

Second, we'll consider that decision metrics may equate to owl:DatatypeProperty components and look at how SPARQL can actually represent an assessment of the options based on the metrics. Finally we'll consider how enhancements like weighting the metrics might be addressed and some issues which came up during the exploration.

Bruce: Why is there so much interest in timelines and maps?

Eva: Linked data, if attached to an ontology, it's minimal, so not much inferencing possible. The value is that you can combine the datasets together.
... The ontologies are very, very simple. People are exploring the fact that they can link a piece of data to a time or space. To talk about ontologies as such,
... this isn't the best source for reasoning.

Jeff: Regarding using open data sets, for metrics it seems datatype properties are "metrics", does it make sense have owl:DatatypeProperty subclass the Metric class?

Eva: I have to think about that, cause the question is what do you want to do with this?

Jeff: Give me all the metrics for the dataset is an example

Eva: But interesting meta-modeling things, so if we want to do reasoning, we have the constraints of the OWL language and I'm not sure how to represent those things

Properties of properties?

Eva: Problem is that you can't reason with that meta-modeling.
... You see the meta-model there. But you are not allowed to do the same step again to reason, cause then you are in OWL full.
... But there are many ways around this.

don: I will add out to the wiki and move some more things forward

Thanks everyone for joining the call and participating in this effort.

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2010/06/24 15:01:26 $

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: Bruce Don Eva Jeff ScribeNick
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_7_Agenda
Got date from IRC log name: 24 Jun 2010
Guessing minutes URL: http://www.w3.org/2010/06/24-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]