02 Sep 2010


See also: IRC log


Jeff Waters and Don McGarry
Jeff Waters


<scribe> Scribe: Jeff Waters

<scribe> ScribeNick:jeffw

Welcome to the 12th meeting of the Decisions Incubator! We’ve just submitted a position paper to the Workshop on Ontology Design Patterns (WOP 2010) describing our Incubator and its efforts. This is an important piece of documentation and many thanks to Eva for her efforts and role as editor for this.

I see that we are at an interesting part of our work, where we have an initial development on our design coupled with some interesting work on a tool and an interesting domain to explore it with emergency management. The recent paper describes our path forward and we have a few topics today covering the range of items ahead for us

eblomqvi: Agree that we are able to collect our thoughts with the development of this paper. I agree with your agenda for today.

Update or Questions Re Core Data Model

jeffw: On my part, I was looking at your example, is this something you could walk us through?

eblomqvi: I will look at the example in my tool and we can walk through it. I haven't had a chance to do any more on the core model
... I used Topbraid Composer out of habit and it has some other functionality and then the Neon Toolkit has the nice support for the patterns.

jeffw: While we're thinking, I was looking at the outline on our wiki of our final report, perhaps we can use what we did on this paper in some ways to fill in our outline

eblomqvi: Ok I want to buy some product, and I have some options, shop1 or 2 or 3 and those are the options and my question, this is what I mentioned last time, we have the information realization, some concrete piece of text
... which is the realization of the question. I don't know if we need the distinction, but in this example I did make the distinction, so the shops have the realizedby property which associates them with some shop data
... Shop1 is an abstract option, then the realization is the entry from this linked data set which is the concrete thing we are using as the option
... In this case, it's not really the shop, it's the info that I buy something from shopX, that is my option,
... What I wanted to get at here, the thing, we want to have a model that can capture when we have a sentence as an option, "I buy the product at Shop A", not machine understandable, then we have the case where we have the data, some entries in a dataset or text describing the option
... The thought is to capture both, where one is an extension of the other one. So you always have the abstract case but you can extend it with additional detail. But I'm thinking that the realization may not be the correct relationship.
... Do you see that it might be nice to have different levels of options ways to express it?

jeffw: sounds like a thing and its description?

eblomqvi: Yes, maybe that's the better distinction

jeffw: I was thinking that the concept/realization distinction is also important, so you could say what computer should I buy, and then at the concept level you might say company A, B or C but then you need to get to a specific model

eblomqvi: Also my idea was that a similar thing for the question, the question could also have a more detailed description, if you wanted to do an automatic assessment, you might want a Statement with Variable pattern, if you want to know what or when, then similarly you could do that for the options
... But at moment in the example, the question is just a sentence. Then I have this criteria, so there were two, one was called buyingCriteriaPrice and maybe one criteria applies to one option or more, so I want the best price from the one closest to me as an example

jeffw: So you had two criteria, one was buyingCriteriaPrice and one was buyingCriteriaDistance

eblomqvi: So in the example, the distance criteria applies to all three options but the price criteria applies only to two of the options, so one shop I don't care how much it costs or I would expect price not to be an issue there. So you can have different criteria for different options.

jeffw: But this is great, a wonderful way for us to do the design and then paint on the canvas to generate these types of examples and then go back and repaint, so to speak

eblomqvi: I can go back over next couple weeks and we can do some more examples. And then we could have some text to match the model.

jeffw: I like this idea of annotating our model with text in our wiki for example to educate folks and to work from small example to larger but also to explain components of larger example
... Also the concept of keeping all this modular so it can be reused

Suggested Process for Exploring and Validating Models

jeffw: basic idea is that we want to try to evolve our decision model by trial and we would like to do this with a tool that could speak this format and be used for decision-making, and we'd also like some sort of infrastructure support, like a repository and other components, in the form of a what I can all interoperability testbed.
... And then ideally, we could interact with the testbed through the tool and perhaps have mobile applications on our smartphones or pad/tablet computers
... to make it easier and funner to utilize the tools and then capitalize on the linked data sets and show how we can use those to drive the tool.
... Then what's intersting about that is we could then tie in any other data or even data that we self-generate, and all we need to do is get it in the linked data format and then we can drive the tool with that data.
... (described the concept combining the data model with the tool with the open linked data format with the interoperability testbed with the mobile apps)

eblomqvi: We also have a tool that could be useful, for example alignment rules, we've used on xml and database schema, when you convert the data from those formats to RDF then you can with this tool based on the rules you can align it with an ontology that you want to describe it with.
... So if we have data in our decision format, then we could link to the decisions with the underlying ontology and I will investigate the state of the tool or web services

Conversion of Emergency Management Data into Linked Data format.

jeffw: So just to touch upon a couple of our remaining agenda topics, since we are running out of time today, there is the possibility of taking some of the types of emergency management data (alerting, hospital availability, situation reporting) and converting into RDF in a similar format to that used in the Open Linked Data sets.
... We've already done some of this and more can be done. This data could then be used with any of our tools. Perhaps next time, a good agenda item would be to invite some of our tool developers from our different members to discuss the status and current capability of the tools.
... We could also then summarize our current thinking and format for them and discuss a common iterative schedule.
... I very much appreciate the idea of using the CNR tool you mentioned, Eva, and it makes an important point to show a variety of tools utilizing the format to confirm its relevance and usability.
... This might be a good opportunity to demonstrate the tools and capabilities in this domain of decision-making, and provide additional relevance of our work to other components of our organizations.

Schedule and Future Papers

jeffw: We're out of time for today, but I would suggest that we outline our iterative development process with our tool developers next time. The results of our ongoing work can then be the update to the position paper recently completed.
... Thanks for attending.

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2010/09/02 15:21:37 $

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: ScribeNick eblomqvi jeffw
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_12_Agenda
Got date from IRC log name: 02 Sep 2010
Guessing minutes URL: http://www.w3.org/2010/09/02-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]