Minutes of SOCoP Meeting Wednesday Jan. 20, 2009, from 11:00 - 12:10 EDT (2IZS)
Attendees: Nancy Wiegand (UW), Gary Berg-Cross (Knowledge Strategies),John Moeller (NG), Naijun Zhou (UMD), Xavier Lopez (Oracle), Todd Pehle ( Orbis Technology ) , Brand Niemann (EPA), Dave Kolas (BBN Tech) Josh Lieberman (Traverse Technologies) (2IZT)
Summary of Topical Discussion (2IZU)
1. OGC RFC specification (OGC 09-157) defining spatial queries using SPARQL Initiative (2IZV)
Xavier Lopez (Oracle) provides an update on this for the community and those interested in participating. A copy of the recent Spec can be seen at: (2IZW)
http://ontolog.cim3.net/cgi-bin/wiki.pl?OgcDraft_Candidate_Standard (2IZX)
Xavier explained that development of the draft Spec, although currently a ¡§sketch¡¨ was a consolidated effort within OGC. The strategy is to leverage well defined OGC models by expanded then to RDF. OGC is trying to bring the best minds together on this effort. The WG has used many sources that are well defined not just Oracle Spatial. So while it resides within the OGC community they taking comments from within and without. Thus the offer for SOCoP members to participate still stands and they are still looking for other participants/members to work on the Spec. As an example John Goodwin at the Ordnance Survey did comment and provided sample data. (2IZY)
There is no specific timelines going forward except that they might like to advance to next level review by March and are considering a call date with people who express interest in Feb. but this might be pushed out to March. The SOCoP members thought that they had enough to go forward for now. The next OGC Technical Committee meeting will be held March 8-12 in Frascati, Italy, hosted and sponsored by ESA/ESRIN. The Spec will be discussed there. (2IZZ)
Josh Lieberman described 3 aspects of the work on the Spec. (2J00)
a. Encoding - Includes known encodings b. Behavior ¡V This discussed such things as how to include filter object in standard a or b etc. It is harder to understand s what the behavior or Spec will be with these extensions and how it can solve problems for them. It is hard to form requirements for how specified services will behave with a particular requirement. (2J01)
c. Implementation ¡V The last part is how it will work when implemented with other people¡¦s encoding. (2J02)
Dave Kolas asked what was the best way to participate and respond to the Spec. Xavier said that the easy way is to forward comment example of a query by Spec section. Commenters can also point out where definitions are weak or where more work needs to be done. The Editor, John Perry, wants a normative section and then a section with examples. (2J03)
Dalia (USGS) said that she would like to participate. She has asked one of her Comp students to do a search of the Specification against their (TNM) spatial relations. The student may be finished with this by the end of Feb. (2J04)
2. GEOSS Phase3 pilot (2J05)
Josh Lieberman provided an overview of GEOSS Phase3 pilot which includes registration and mediation/mapping on ontologies. A copy of the summary he provided is available at http://ontolog.cim3.net/cgi-bin/wiki.pl?GeossPilots (2J06)
He also noted that they aresSoliciting proposals at Phase3 pilot the end of the month. (2J07)
3. Update on USGS integrated National Map (TNM) (Dalia Veranka). (2J08)
Dalia noted that the work hasn¡¦t gone as fast as they wish. They are exporting from oracle to triple stores, but are still working on that and GeoNames. She thought that finish this part in 4-6 weeks. Lynn Usery is working on the raster data as a related effort. (2J09)
Dalia noted that he U.S. Geological Survey¡¦s Volunteered Geographic Information Workshop was held in Herndon, VA January 12 ¡V 13, 2010 to gather information on the potential use of volunteered geographic information as part of The National Map. The workshop reflecting the influence of increasing amount of metadata about user-generated content from Web 2.0 sites which is becoming available as Linked Data. To this end the volunteered geographic information workshop at allowed users to generate content. The web page for this geographic information workshop listing agenda and speakers is at: http://cegis.usgs.gov/vgi/ (2J0A)
The workshop was initiated from Production side of TNM and the meeting was encouraging. Dalia mentioned that Open Street Map (OSM) effort. They use volunteer surveys who work on foot, bicycle or in a car. Map data is usually collected using a GPS unit, although this isn't strictly necessary if an area has already been traced from satellite imagery. Qualified people are involved to provide good lineage info on who generates it etc. OSM takes everything and is not filtered like USGS to make a trusted product. See http://www.openstreetmap.org/ (2J0B)
Gary asked a question on Open Gov and how USGS is proceeding with this initiative. Dalia noted they had a briefing, and summaries will be available. (2J0C)
4. Tools for Developing a Geospatial Service ¡V update from action items (2J0D)
Gary introduced the topic and suggested that the discussion and work would be in 5 parts started with a Use Case. This builds on a three-layer client/server architecture: U The data layer U The business logic/application layer U The view (GUI/presentation) layer (2J0E)
To this SW we add a Server HW layer, since the data and SW has to reside somewhere. These 5 parts are listed below with notes as they were discussed. To move forward we need to identify a plan of action and resources for each. (2J0F)
a. Use Case ¡V we currently do not have an overarching use case which may be part be driven by The National Map data and what other data is publically available. We need some one (or group) to take the lead in this and make a Strawman suggestion. Luis Bermudez might be able to make suggestions, but Dalia should also provide some guidance. (2J0G)
b. Server (HW) Currently the idea is to use a USGS server so the resulting service is available to the public from that source. An alternative suggested was to use the Google Cloud. If the SOCoP INTEROP proposal is funded there may be a server available as part of that effort. (2J0H)
c. Data Sources, were discussed extensively at the last meeting and these included: i. atmospheric: http://mesowest.utah.edu/index.html ii. geo names http://www.geonames.org/ iii. oceans iv. DBPedia v. USGS Maps vi. Data.gov See http://ontolog.cim3.net/cgi-bin/wiki.pl?SocopMeetingpage_Minutes_12_19_2009 for information about data sources and the service idea. (2J0I)
d. Application - Todd Pehle was not able to attend but provided these ideas on using a Sesame SAIL that would be created is part of a Sesame Server - basically it¡¦s a SPARQL endpoint. It also serves as an API for external Sesame application code. Todd thought that he could participate flexibly and either help provide a server-based sparql endpoint service and/or could write some the application logic, too. There may be an issue with licensing Oracle to support this. (2J0J)
He thought that it ¡§would be cool¡¨ if the demo included more than 1 sparql service or distributed RDF / RDFa, etc. This would make it feel more like a Semantic Web/Linked Data application than a ¡¥semantic technology¡¦ demonstration. An alternative to this approach is to use a Google App Engine, but the group did not discuss this option further. (2J0K)
e. GUI layer Naijun had some suggestions. There are research and tools developed by the HCIL at University of Maryland, and it is likely to involve them but he would need to ask first. Naijun Zhou has already used SpaceTree developed by Prof. Ben Bederson's group for his ontology and semantics work. But these tools have no implications of geographical locations if used for spatial ontology and semantics." (2J0L)
Penn State has a world famous visualization group: GEOVISTA from geography and information science.: http://www.geovista.psu.edu/. Some work has been done for ontology, but curiously there seems to be not work linking the visualization for geospatial apps and ontology apps. Josh noted that it is a curious gap. There is ontology visualization and map viz, but nothing combining these. He was intrigued by the idea of adding it as a data type to open layers. (2J0M)
Other tools were discussed. ¡§Palantir¡¨ is being used by NGA and other agencies. TNM is using this and the new national map is a Palentir viewer. Progress but still has limitations on querying data from the viewer and people need to look at it closer. This is more of a traditional map viewer, with many layers. Xavier noted that we may be talking about 2 diff things. One is commercial product to display data. (2J0N)
Idea for moving forward. The suggestion was that we¡¦ll see how linked data plays out and what issues they run into to help us move forward. (2J0O)
5. Status of Open Ontology Repository efforts relevant to SOCoP (Mike Dean) (2J0P)
Mike was not able to attend, but provided some links so people can start now using the sandbox at [1]. At some point soon it will move to [2]. Mike is hoping to populate it with the ontologies from the old sandbox [1] shortly. (2J0Q)
People can also create your own private repository. BBN recently did this, but it took a while to install all the software dependencies. (2J0R)
[1] http://oor-01f.cim3.net (2J0S)
[2] http://oor-01.cim3.net (2J0T)
6. Nancy Wiegand indicated there was no news on the status of our NSF INTEROP proposal (2J0U)
7. Because of jury duty Todd Pehle was not available to update the group on the next VOCamp plan. However by mail he noted that ¡§since SOCoP is developing a real use case and demonstration, he may use this SOCoP scenario to help drive the development of our mini geo-onts in the neogeo semweb vocabs google group. (2J0V)
There being no Other Items/New Business the next meeting was targeted for Feb. 17, 2010. (2J0W)