Decision Mtg 10 Agenda

Decision Meeting #9 Goal & Agenda

Background

The dial-in information for this week's meeting is here. The Decision Incubator Activity is intended to explore the use cases and design of a Decision Representation format (e.g. in XML, RDF, OWL…) available on the wiki along with its use by participants in testbed settings for experiments and exercises. Each meeting will be an opportunity to summarize progress & contributions to the wiki, review next steps and action items, and to review the current representation format.

Summary from the Last Meeting

At the last meeting (see minutes), we reviewed the status of the Provenance Incubator work, Piotr Nowara provided an overview of his decision ontology efforts, as did Marion Ceruti, a concept for converting emergency management data into RDF was discussed, as well as a best practice for handling reification using named graphs.

Regarding the Provenance Incubator, they are doing excellent work and have a nice report on requirements here. Their next major report on a review of the current technologies and identification of gaps is underway. Decisions are relevant as a source of provenance information. The work performed in the Provenance Incubator is of great significance and we will continue to monitor the progress and consider the implications for our efforts. Also noted was the work on the Open Provenance Model (OPM) which motivated some of the Provenance Incubator effort. The OPM was intended to address the provenance of a wide range of items, including decisions.

Regarding Piotr Nowara's ontology work, there are two use cases driving his efforts, one is describing decisions from formal legal, industry standards and the second creating decisions based on actual experience. Piotr has his initial ontology work available here. Marion Ceruti mentioned her related work on the ontology of cognition, the interaction of language and culture, where a key enabling concept is recursive subsets. Everyone appreciated Piotr's and Marion's ideas, and they will continue to contribute and share their perspectives so we can benefit from their knowledge and efforts.

Regarding emergency management data, both Jeff and Don are participating in projects and interoperability testbed activities intended to improve sharing of emergency management information utilizing standards. The OASIS Emergency Date Exchange Language is a family of standards which are currently being used to represent alerts, hospital availability, distribution elements, situation reporting, patient tracking and other emerging types of information. One option for exploring improved interoperability is to simultaneously convert these XML formats into RDF, similar to what is being done on the Open Linked Data initiative. Initial efforts are being explored and Jeff and Don will keep the group informed of progress. This data could then be more easily utilized for decision-making using the patterns being modeled in this incubator.

Regarding best practices, the issue of reification (where for example, the decision "question" could be a reference to a "triple") is being handled in several systems by using quad stores instead of triple stores, where the extra component beyond subject/predicate/object is a graph id. Since a graph can consist of any number of triples, including only a single triple, the graph uri can be used as a unique id for efficient and direct referencing overcoming some of the perceived weaknesses of traditional RDF reification. This looks like an appropriate solution for some of the decision reification needs mentioned.

Goal for this Meeting

At this tenth meeting of the Decision Incubator, we are approximately half way through our incubator's one year lifespan. It is a good time to take stock and recommend a schedule for completion of the work. Four main agenda items will be considered. First, the full breakout of a decision format into patterns will be suggested. The remaining work for exploring the requirements through doing, i.e. eXtreme design, can then be scheduled as the application or implementation of these patterns. Second, to address some immediate needs, a simple XML format will be revisited. The suggestion is that the XML format can be converted into an initially simple RDF (and such a conversion standardized via GRDDL) and that this will provide a mechanism to begin using the emerging decision format while the conversion is upgraded to the full pattern-based design as it reaches completion. Third, a recommendation to begin filling out the final report outline will be considered. The goal is to concisely summarize the discussion and exploration of the incubator work in a fashion that can be easily understood and appreciated by others. Fourth, some implications for routing will be considered for both decisions and interesting triples.

Agenda

The following agenda items will be discussed at the next meeting. The topic, time, summary and discussion leader are provided for each agenda item.

(0) Review of Status, Goals & Agenda Overview (5 min, Jeff)

Use of our component models will be mentioned and suggested goals for next few weeks.

(1) A Decision Format as a Set of Patterns (15 min, Jeff/Eva)

Jeff will suggest that a mapping of decision components to implied patterns can be performed to generate a relatively complete set of patterns to encapsulate the desired decision format. This list of patterns can then be described, implemented (or reused) and integrated according to a schedule. Decision components have already been identified but the implied common patterns in those components can be more clearly elucidated and more directly understood and scheduled.

(2) A Simple XML format for Conversion as Open Linked Data (10 minutes, Jeff)

While the design work continues, there is a need to do some initial exploration of the decision components integrated in at least a simple xml format. This format can then be converted into RDF in the same fashion as is done in the open linked data initiative, where a fairly straightforward and simple conversion is performed in many cases. This will help provide an example of going from A-to-Z with raw data, conversion, query and exploration. The conversion into the more expressive designed pattern-based ontologies can be a natural extension of this near-term effort.

(3) A Way Ahead via Interoperability Testbed and Open Linked Data Example

To complete the incubator activity in a fairly robust manner, the following steps can be considered: (1) Do items 1 and 2 above in parallel; (2) Implement in testbed an RDF datastore and query engine as basic infrastructure back-end capability;(2) Instrument an application to utilize the format from item 2 above (updating the format as item 1 proceeds) utilizing the testbed; (3) Utilize the application in some events/exercises/daily activities to explore the format; (4) Proceed with the documentation effort in parallel.

(4) Implications for Routing (10 min, Jeff/Don)

Jeff and Don are collaborating with Solace Systems for improved routing of emergency information. Routing of decisions is a natural extension of the current work. In addition, the concept of enabling improved interoperability through breaking apart stove-piped XML Schema-oriented datasets into flexible, RDF-oriented datastores offers an opportunity to consider the implications for routing. Rather than simply route entire messages sets, what about routing triples of interest? Participants can briefly consider this topic and additional follow-up is possible, including exploration in a virtual interoperability testbed.