Minutes of the Wednesday March 17th 2010 SOCoP Meeting (11:00 - 12:10 EDT) (2CRC)
Participants: Eric Wolf (USGS), Mike Dean (BBN Tech) Dave Kolas(BBN Tech), Nancy Wiegand (UW), Dalia Vernaka (USGS), Gary Berg-Cross (Knowledge Strategies), Todd Pehle (Orbis Technology), Dave Kolas (BBN Tech) Carl Diegert (Sandia Labs), Ron Fresne (SRI/ National counter-terrorism centers) (2CRD)
These minutes can be updated and corrected here. This includes some additional links to topics like the 2010 Workplan. Over time some additional links may be added on related material and sections on topics created that accumulate and organize material from successive meetings. (2CRE)
Following introductions the agenda included 6 main topics: (2CRF)
1. Update on USGS integrated National Map (TNM) data becoming available. (2CS8)
Eric Wolf summarized the discussion with Todd on technical details such as the use of the USGS server and how to structure shape files. They might add some features to their such as bounding boxes. These would be of use when Todd extends SAIL. Steve Hildebrand is working on the Oracle licensing for an outward facing application is different from internal use, but as a demo it still may be free. (2CRG)
Dalia was looking at the proposed SOCoP work plan and is thrilled to be working with SOCoP. She noted that it is important to keep in mind the eventual goal to develop improved National Map ontologies and put this in an OOR. This is based on ideas of what the OOR would accept. One question she has was whether the Repository should accept individual data (instances) or just the ontological classes. Mike and Gary offered some ideas on this. Mike noted that the original focus is on the classes rather than a storage of massive amounts of instance data. Gary suggested that some sample data might be stored as part of the annotation. This might be particularly interesting to illustrate mapping and serve as a half way point for semantic mediators. (2CRH)
2. Discussion of several subparts of Workplan for 2010. (2CS4)
a. Planning for a good presentation/demo for Tech Days was proceeding per the notes on the demo below. (2CRI)
b. Todd Pehle discussed his efforts to minimally implement a geo/spatial sparql using SAIL ((Storage And Inference Layer). He has been collaborating with Xavier on this as well as using Oracles Sesame adapter to maps TripleSet data model. Todd is using the notion of query hints to understand what is happening under the hood. He believes that he has a way of supporting most of the operators to make it look like geo SPARQL. This would be implemented in a Sesame stackable SAIL (sitting on top of Oracle Sesame Adaptor) that would rewrite incoming geoSPARQL with the necessary query hints (thus abstracting from the client). In addition, it will encapsulate the ability to load incoming RDF (containing geometries) into the underlying Oracle Spatial tables. This may be implemented in the next few weeks at which time he will hook back up with Eric at USGS to move demo along. Once this is done in the next month we can move closer to a Use Case and ideas for developing a GUI client using OpenLayers. Both will follow work on the query. (2CRJ)
Todd is in discussion with Mike & Dave about the possibility of using Parliament to host other (non-USGS) data. He hopes this can this come together soon in time for the demo at NGA Tech Days. (2CRK)
Eric has talked Todd to attend WhereCamp (directly after Where 2.0 conference). This at the end of March. So they will hopefully have a chance to integrate the demo at that time . Also, Eric said hell be in D.C. in mid-April so the 2 should have yet another opportunity to get the demo going. (2CRL)
c. Update on the OGC Geosemantics session at the Frascati TC Josh provided an update in GeoSemantic meeting which he thought was a juncture somewhere between IE, to represent an query spatial data, and what Oracle and other people are doing to have a geospatial SPARQL. Besides these two there was also a 3rd proposal by an Athens group for a geospatial SPARQL (stRDF and stSPARQL). There are Italian groups working on semantic broker components but it is unclear what SPARQL approach they might adopt. Dave Kolas has had a look at the Athens group slides and doesn't think it's all that different of an approach from previous ones. They had proposed some additional constraints on the RDF. Dave asked about this and as part of Joshs response he noted that they had added a time term to the triples. It wasnt clear why this was although Gary suggested that triples are arbitrary and that adding time to a 3D world gives a 4D world in which events can happed and objects are located in time and move. (2CRM)
Todd asked about the plans for the geoSPARQL stardard and when people might be able to do standard implementations. The issue for him is that standard might change and he is trying to implement a geoSPARQL in the near term. (2CRN)
Josh thought that the meeting and participation with multiple approaches provided good momentum. (2CRO)
In follow up Gary asked if slides from the Greek group might be available. Josh noted they are available to OGC members but would look into posting them at an intermediate site, which we can link to. We may post information about such things to the SOCoP forum. Josh and Gary will talk further about how to handle this. (2CRP)
d. Plans for June the Geosemantics DWG see http://www.opengeospatial.org/event/1006tc for information on the meeting. We would like to put together a draft for the June meeting and hope that our Demo is ready to feature this at the meeting. (2CRQ)
e. Further development and collaboration on a Reference Model (RM) (2CRR)
Interest in and the value of a RM was sprinkled throughout the meeting. For example one issue discussed with Todd concerned the use of semantic mediation to create data federation and handle cognitive issues. There are different ways to annotate and encode interface standards which suggests the need of a framework to do this. Josh suggested that we might first build a Reference Model. This may be an abstract pattern which addresses a semantic use case that employs OGC standards. The Framework would address how do one encodes spatial data in OWL and queries it, does semantic operations, response translation etc. Such work and its formalization would move SOCoPs Reference Model forward. (2CRS)
Another relevant activity, mentioned by Josh, was the Vocabulary and semantics WG within the GEOS pilot. There are several groups involved (e.g. MMI and one from Tokyo). The idea is to support vocabulary registration, mapping and query expansion. Discussion of this is on a Google Groups site. Josh will provide a link. The site is open to view but editing is limited. (2CRT)
The first planned step is to build in a voc registration process as part of the work. The WG made up of what Josh described as capable people is using SKOS. The MMI folks, for example, are building their own OOR. (2CRU)
3. Efforts to populate the Open Ontology Repository for SOCoP (2CS5)
Mike Dean noted that there was more work on federation to support the repository and will offer an OOR presentation at the semantic Tech event in June. Gary Berg-Cross had been in touch with John Goodwin to discuss a schedule to add further Ordnance Survey ontologies. These are currently being updated, so it will be several weeks or a month before the new versions are available. In the meantime Gary is contacting others for candidate ontologies for the OOR. One example is SOUPA (Standard Ontology for Ubiquitous and Pervasive Applications) developed at UMBC. (2CRV)
Nancy had a question had a question whether this work represented a new instance of the OOR or a sandbox. Mik replied that if we get funded by NSF it would be a new instance with more gate keeping funding and extend the BioPortal categories. Currently it is all in the sandbox. MMI has their own instance and we would like to get some of their extensions. (2CRW)
4. The 2010 Ontology Summit At NIST (2CS6)
Gary noted that the 5th Annual Ontology Summit on Creating the Ontologists of the Future was held march 15-16. Information and links including the resulting communiqué are available at: (2CRX)
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2010. (2CRY)
A driver for the meeting was the group observation that projects centered on ontology technology are now being advanced by governments and by scientific and industrial organizations. The result is a growing need for ontology expertise which in turn means a need for well trained ontologists. The 2010 Ontology Summit explored strategies to address this need in terms of: (2CRZ)
- curriculum, (2CS9)
- establishment of new career tracks, (2CSA)
- the role of ontology support organizations and funding agencies, as well as (2CSB)
- training in the analysis and comparison of methodologies for designing, maintaining, implementing, testing and applying ontologies and associated tools and resources. (2CS0)
5. Nancy Wiegand reported that there was no change to the status of NSF INTEROP proposal. (2CS1)
6. Update on use of Ontolog SOCoP Forum for discussions (2CS7)
Gary noted that the forum may be used to discuss some of ongoing issues and provide update information such as links to the Geosemantics WG. Mike thought that the Forum could be useful and encouraged its further use. (2CS2)
There being no other Items or New Business the meeting wrapped-up and Gary suggested April 14 or 21 as possible dates for the next meeting. (2CS3)