Minutes of Dec, 2009 SOCoP Meeting (2IZ0)
Wednesday Dec. 16, 2009, from 11:00 - 12:10 EDT (2IZ1)
Attendees: Luis Nancy Wiegand (UW), Gary Berg-Cross (Knowledge Strategies), Kevin Backe Army Corps of Eng.), Peter Yim (Ontolog), Naijun Zhou (UMD), Luis Bermudez (SURA), Brand Niemann (EPA) who popped in at the end of the session. (2IZ2)
There being no change in status of NSF INTEROP proposal effort as reported by Nancy Wiegand, the meeting largely served as an update on action items from the Nov. 12th workshop. The main focus was on follow up to ideas on the Workshop breakout on an Integrated Demo. At this meeting several people thought that publically available integrated National Map (TNM) data becoming available (between now and 2015) provided a new demonstration opportunity. For example, he might provide a good platform for testing within the Semantic Web paradigm of linked data. More than that it might provide an opportunity to build a publically available service and deploy this with supporting tools. Three main topics were suggested by Luis to advance these ideas: generally available data, how to handle TNM raster data and client SW to leverage for the development needed. (2IZ3)
For more information, contact the SOCoP Executive Secretary - Dr. Gary Berg-Cross via email to: gbergcross [at] gmail.com (2IZ4)
General Data Sources Luis, started listing several sources which eventually had 6 types/sources (2IZ5)
1. atmospheric: http://mesowest.utah.edu/index.html 2. geo names http://www.geonames.org/ 3. oceans 4. DBPedia 5. USGS Maps 6. Data.gov (2IZ6)
Data.gov has recently moved to use semantic web technologies to expose metadata and structure of information to link data on sites like recovery.gov with information shared through data.gov. (2IZ7)
See http://data-gov.tw.rpi.edu/wiki/Data.gov_Catalog for the catalog providing a listing of data.gov datasets converted into RDF http://data-gov.tw.rpi.edu/. The site notes there are 1081 datasets published at data.gov (see listing and example is the 2005 Toxics Release Inventory National data file of all US States and Territories). The conversion has generated 116 RDF datasets contributing 5074936168 triples. (2IZ8)
Issue on Raster Data (2IZ9)
Dalia had sent an email on the status of the National Map triple stores and the CEGIS server access. There is progress and both, but the end of the year has been too busy to set a date for the data release announcement. One problem is that 3 of the 8 data layer 'themes' are raster data and Dalia was not aware of how to represent raster data as triple stores. This is important because the raster layers themes are some of the most widely-used public data (land use, elevation, and ortho-imagery). (2IZA)
One idea is to use WMS and describe the images of interest with RDF. Alternatively, Naijn suggested that we dont need to process/covert the whole raster data set to RDF. We can finesse the issue because there is a whole image represented by the data, so just link to it and download and use it as an overlay. It serves as background image info. (2IZB)
Then if we need a new points we go to vector. We save them as x and y and plug in as a first stage of work following the MMI style of work. (2IZC)
Leverage Client SW (2IZD)
This was the longest discussion launched by the question of who would generate a use case, which was converted to the question of Who is going to develop the client? One related question was a tentative schedule, with Gary suggesting that the model from prior years was to look to an early one in May and a more substantial one in Nov. The Marine Metadata Interoperability (MMI) Project was mentioned as a possible example, but Luis noted hat it is hard to move forward. Instead he suggested a project at RPI that will end in 2 weeks, which may have a useable client and may want to demo it. Luis took an action to contact Peter Fox (http://tw.rpi.edu/wiki/Peter_Fox) at RPI. (2IZE)
Naijun might be able to commit to developing a client if we receive NSF funding for our proposal. This means waiting till Jan or Feb to see if we are funded. (2IZF)
Another idea was to see if OGC support is available. Xavier Lopez (Oracle) sent a message to inform the SOCoP community that a draft RFC specification (OGC 09-157) defining spatial queries using SPARQL has been introduced to the Open Geospatial Consortium (OGC). Xavier invited participation from OGC members (and other interested parties) to support the drafting and completion of this specification concerning development of a spatial language component to SPARQL as part of Pilot Project 3 was announced last week at the OGC semantic meeting Central to this SPARQL definition will be a proposed ontology for spatial relations (e.g. within, crosses). We might participate on a task for them aligned to this project. Naijun, asked if this work is about spatial relations, does it fit into the 3 areas on relations (e.g. foundational ontologies, qualitative calculi, linguistics) as well as geometrical) that John Bateman discussed at the workshop. People will have to study the announcement and spec to answer this question. (2IZG)
John Moeller should query Mark R. about how the community might participate in this. Kevin asked if the SOCoP might collectively be interested in putting together a proposal to participate in this phase by March. We would be putting some resources in and formalizing this as part of our proposal. (2IZH)
One issue was that OGC may want some people to become members to work on the project. We might get involved individually or collectively as SOCoP members. We might check also check with Joshua Lieberman. (2IZI)
Someone thought that we might be able to leverage some of the GEOSS work by GMU Doub Niebert mentioned at the workshop. We might be able to align to OGC leading a Phase 3 pilot GEOSS project. Gary will try to follow up on that and see if they will help us leverage their demo. (2IZJ)
Luis explained that many groups (10-15 groups) are participating in GEOSS (http://www.earthobservations.org/) and registering services. Kevin said added it was charted out of white house and NASA involved with many earth observing groups in other countries. Luis, for example, was part of the GEOSS data group. (2IZK)
Luis was asked in the client might shows how semantic mediation is used and thought that this could be done with new or adapted SW. Some thought that we could use the codebase from Intercoastal work. Others asked if Mashups possible? The answer seems to be yes, but come down to who is going to do either build a client of adapt one. Even if we do it with Google it would still take time. We can do lots of things, but the UI is critical and who would do this? (But see Brands slides on services in the cloud - see http://federalcloudcomputing.wik.is/@api/deki/files/132/=BrandNiemann12152009.ppt for a discussion of A Semantic Cloud Computing Desktop / Mobile Apps with Linked Open Data that consists of: database of "things" referenced by URL's A free Wiki (Deki Express) and a semantic publishing environment that supports use on Mobile Apps (e.g. iPhone) and Linked Open Data through MindTouch Extensions (e.g. App Catalog), conversion of the MySQL database to an RDF triple-store (e.g. DBpedia), etc. (2IZL)
It seems that there should be tools within the SOCoP community usable to provide a client for a demo. Northrop Grumman has one (e.g. Scott Fairgrieves PULSENet, and others include Geo-Connections, Com-Pusult, Cc-Mc and Smart-Bay). (2IZM)
Nancy noted that she is hiring some undergrads (but may not have RDF experience) . So she might coordinate if we have something to coordinate. What might they work on? They might help with the OGC proposal these are they types of maps we have available from USGS etc. We want to use TNM and need to know what data will be available as triple stores. One idea was to leverage OOR work through a registry of ontologies. (2IZN)
Kevin suggested that the next conversation we should have is on how we want to participate and what resources to use. (2IZO)
There being no Other Items/New Business the next meeting was targeted for Jan. 20, 2010. (2IZP)