From VoCamp Wiki
Jump to: navigation, search

Mailing list[edit]

Some workgroups requested that a mail group be set up to continue discussion.

The link below is for the Gov Linked Data WG started on the first day.



GeoVoCampDC2011 was a follow-up to GeoVoCampSouthampton2011, but was also an open opportunity for people to discuss topics they are interested in. Some of the organizers have a general interest in linked data. There is also an interest on near-term geographic vocabularies for Linked Data which could be a wide use. Another interest is in geo-spatial issues to support the NSF SOCoP INTEROP project. SOCoP (Spatial Ontology Community of Practice) INTEROP work is an effort to develop a collaborative, interdisciplinary network to support geospatial data sharing, integration and interoperability. Current community efforts are starting to develop linking ontologies to support better interoperable for Land Use. We may discuss Land Forms and Hydrology as time permits with overlap to Marine/Ocean and other Earth Science data. See

The Geo Web offers a massive amount of structured information that could greatly enhance the Web of Linked Data. However, to begin the convergence of the Semantic Web and Geo Web a small set of core vocabularies is needed. Once constructed, the "Web of Linked Locations" will make the Web appear as a massive distributed interoperable spatial database that interlinks locations with other locations, events, sensors, people and everything in between.

Some breakouts at VoCampDC2011 will seek to continue the initial efforts of the W3C Geo XG, NeoGeoVoCamp, GeoSPARQL initiative at the Open Geospatial Consortium (OGC) as well as the most recent GeoVoCampSouthampton2011.We hope to blend participants from the Semantic Web as well as the Geo Web. If you are interested in the spatial projection of the Web of Linked Data, help us pioneer the new spatial frontier!

-Todd and Gary

Besides, isn't it time we put the "geo graph" into "geography"?

Please list any geo vocabulary topics you'd like to discuss under the Topics section below.


3rd - 4th June 2011

Day 1 (Friday 03 June 2011)[edit]

09:00am - 5:00pm

Day 2 (Saturday 04 June 2011)[edit]

09:30 - 4:00pm


Day 1:[edit]



  • 12:30 - 13:30pm : Lunch
  • 13:30 - 13:40  : SWEET (Rob Jaskin)
  • 13:40 - 14:15  : Introduction to Web-Based Ontology Engineering (Mike Lang Jr.) and Brief tutorial on using Revelytix ontology tools for sessions ((Mike Lang Jr.)
  • 14:15 - 15:15pm : Vocabulary Breakout Session 1
             GeoSPARQL and RDF:Geometry Breakout Group
             Gov LOD
  • 15:15 - 15:30pm : Socialize/Break
  • 15:30 - 17:00pm : Vocabulary Breakout Session 2


  • Mandatory Meeting at Local Pub : )

Day 2:[edit]


  • 09:30 - 10:00am : Arrive and Make More Friends
  • 10:00 - 10:30am : Decide on Vocabulary Topics for the Day
  • 10:30 - 12:30pm : Vocabulary Breakout Session 3


  • 12:30 - 13:15pm : Lunch
  • 13:15 - 15:00pm : Vocabulary Breakout Session 4
  • 15:00 - 16:00pm : Wrap Up & Next Steps


* W3C basic geo review
* Sweet ontology
* Topography & Landforms Modeling
* Ontology Design Patterns
* Bottom-up Ontologies, Methodologies, Clustering,...
* Triplifying Geodata, Geometry ? where to stop
* Feature Type Ontologies, Places of Interest Ontologies
* Land Use Ontology
* Land Forms - Rivers and Emminances
* Geometry Veocab Review / Southampton Vocamp
* Geonames/ Gazetteers issues (LOD Gazetteers)

More notes at:

GeoSPARQL and RDF:Geometry Breakout Group[edit]


* Different levels of triplification
* No URIs for pixels but for the metadata
* Features and their identity
*  Curve = rdf:List and rdf:first only Point and rdf:rest only' (Curve or rdf:nill)  (
Why would you want to have this? 

Polygons consist of many hundreds or thousands of points.

How would you phrase a SPARQL query for point nr 730947 in a polygon? What is the added reasoning value in contrast to WKT.

* GeoSPARQL is using literals in their RDF representation
* Relations such as 'near' should be computed, this is why we have GIS etc and not hard coded as triples
* RDF:List is not well supported
* Do triple stores ever need to store geometry?  Why?  what is the use-case? 
* Reasoners need persistent identifiers for spatial features, do they need more than this? 
* If one is generalizing landcover classes, then remapping their geometry to create a new generalized map, then how is the geometry to be linked to the classification hierarchy?
* We can make unions of geometries with GeoSPARQL etc all without the above mentioned RDF serialization (e.g., connect all forests that touch)
* Definitional versus inferential, e.g., infer nearness versuch hard-code it, scale, e.g., is an issue. 
  • Without a domain 'near' is meaningless
* SouthOf etc. is already a different case
* Use an ontology design pattern (nested features of a reification with a buffer relation that you would subclass per context)
* Nearness-Ontology-Design-Pattern to be subclassed... reasoning on concrete domains, 
  • you can query and combine different versions of it.
  • the pattern also needs to include a relation to the used metric (e.g., Manhattan)
* Roles for the SPARQL case
* Nearness implemented using a set of rule that define the context
* Ontology Design Patterns can also be made of rules
* The data defines what near is (can be part of a build-in)
* W3C Basic Geo:
  * Use SPARQL construct (going from W3C to GeoSPARQL)
  * Is there a need for model-alignment
  * portable queries
* We should develop best practice guides

Government Linking Data Working Group[edit]

Break out group facilitated by George Thomas of US Health & Human Services and Bernadette Hyland of 3 Round Stones, Inc.

  • 13 people attended the 60 minute break out.
  • The participants included representatives from US Government (US HHS, EPA, Library of Congress), academic, non-profits (Open GeoSpatial - OGC), commercial services and software firms (Orbis Technologies, Deloittes, 3 Round Stones, BBN).
  • We discussed the W3C Government Linked Data Working Group charter, see the main page which was recently launched (May 2011).

Participants who attended VoCampDC 2011 who are interested in attending and are either W3C members or interested in potentially participating as an Invited Expert, are encouraged to seriously review the Deliverables, in the charter and pick one or two items on which they'd be willing to give a 10-minute presentation on during the F2F (either in person or remotely).

The goal is to help people understand that topic. Please send your topic(s) and a few thoughts to the chairs at team-gld-chairs at ASAP. Thanks. Responses are due by Thur 23-June-2011 to help us form the agenda and give us time to give you feedback prior to the June 29-30, 2011 GLD WG F2F, to be held in the Washington DC area.

  • The next GLD WG F2F is slated for end of 2011 or early 2012 in Europe.

Topographic (and Landforms) Ontologies Working Group (IOG)[edit]

Participants: Gary-Berg Cross ( / Gaurav Sinha ( / Ryan Kohl ( / Tri Datta ( / David Mattli (

• After a brief discussion of the potential and need for topographic ontologies (e.g. comparison of terminological Systems discussing Eminences such as GNIS and ENVo), the TOG agreed that a focus on a specific geographic area is necessary to stay grounded in applied ontology and the SoCOP context of interoperability, rather than lose sight of immediate goals in search of unifying, but elusive, foundational domain ontologies.

• The group agreed to pursue the domain of water bodies and use the Hudson River in New York as a use-case inspiration. How do we represent it? The Hudson river not a unitary thing. It has views like a watershed view but also water quality, land level of riverfront, temperature flood zones etc. How to make viewrs relate and any Applications interoperate?

• The discussion then focused on how SoCOP and the NSF-InterOp grant would benefit most from the use-case. It was decided that approaching the semantics formalization for water bodies (chiefly Hudson River) should be done in phases (and perhaps use a mix of top-down and bottom up mathods)s.

         1.	The first phase would be primarily collecting data for the chosen use-case. 
               The USGS should have data as well as several NY groups;
         2.	 the second phase will focus on building a few lightweight ontologies 
                leveraging Hydrology models which will lead to the 
         3.	third phase of engaging a varied group of ontology modeling experts to help improve the design and scope of the 
               lightweight ontologies. 

According to current plans, for the Hudson River use-case, the first phase would involve mainly collecting vocabularies from targeted resources to create a repository of web articles about the Hudson River. People may talk about a river on a web site. We can use USGS and other Maps along with information from GIS, Google maps, texts, tourist books, and Freebase as a start. We can extract primitives from these, and then can use ontologies to help describe this material.

• The targeted approach will allow explicit exploration and represent the ontological commitments of groups with direct or indirect interests in the Hudson River. Example use-cases for the Hudson River discussed included the following:

o water supply, o water quality, o recreation/tourism, o industrial uses, o agricultural uses,( Disasters are motivators. We have Pig farmers vs fisherman) o spiritual connection, o hydrological modeling, o ecological modeling, and o habitat preservation.

Resources for these use cases will be explored by the group. Some of the resources may use a process approach from a Geoscience view, which is not the same as common sense or even a map idea of the water world. This the split between the Science vs. common sense view of phenomena.

As regards to how the ontologies will actually be developed once data sources are identified, there are two possible methods.

1. The first is the traditional method of human driven construction of the ontologies. This may be time intensive.

2. An alternative may be to use some machine learning methods to mine the developed database to identify some ontological primitives as discussed by Jano. We might something like the width of river vs length to define some things. There is also some NLP processing work at Ohio U which might be used to extract concepts by running against such repositories as USGS annotations. Note: This work converges with some Bok2 ideas on classifying knowledge.

In addition to this bottom up and top down mixes may be considered. It is unclear at this stage which method or if both methods will be tested to start developing ontologies for water-bodies and the Hudson River use-case. This decision will have to be made collectively by the group after assessing the effort required, personnel and skills available, and the pragmatics of access to interviewees, books, papers, archives, and quality and volume of online material.

• The ontology building phase will follow these initial efforts to capture domain semantics through the above-mentioned methods. Multiple lightweight ontologies will be constructed and shared widely with a wider community to seek corrections and expanding perspectives. It may also be possible to engage in formal semantic similarity assessments to judge how much overlap and conflict exists between the different use-case contexts. We need the best ontologiests to take on a small part of this problem in the way this has been done in the Biomed field.

• Recognizing that this process should be continual, ontologies will be continually redesigned and improved through collective feedback.

• If and when stable versions are available, we will pursue promoting and publishing about them in different ways to both academic and non-academic information science outlets.

More on Topographic Ontologies… • Ontological Issues for National Topographic Mapping. Cartographica Journal (Special Issue).

• USGS. Building a Semantic Technology for The National Map Research Projects.

• Ordnance Survey (UK) Ontologies.

• Varnaka, Dalia (CEGIS, USGS). A topographic feature taxonomy for a U.S. National Topographic Mapping Ontology.

• Sinha G, Mark D M (2010). Toward a Foundational Ontology of the Landscape. Geographic Information Science, 6th International Conference, GIScience 2010, Zürich, Switzerland, September 14-17, 2010.

• Formal Models of Commonsense Geographic Worlds

• INSPIRE Data Specification on Hydrography


The folks at Library of Congress in Washington D.C. have kindly offered meeting space at their Madison building facility at 101 Independence Avenue, SE Washington, DC 20540. The meeting room is located in the National Digital Library Learning Center within the Madison Building. Once inside the front door of the Madison, head straight towards the big glass doors directly in front of you.

See for a map and directions.

Much thanks to Ed Summers for the help!

This is a public building so you can get in without ID or an escort after 8:30 am.

Send an email to gbergcross at gmail dot com if more details are needed...

Directions: ...see for full details

Closest Metro Stop: Capitol South (orange/blue line)

Exit station using main exit

Walk approx. 2 blocks N on 1st Street SE.

Alternative Metro Stop: Union Station (red line) - .5 miles

A Few Hotels Options:

Capitol Hill Suites has a location that is practically next door

   Capitol Hill Suites

At the other end of the price spectrum, just down the street is the William Penn House:

   William Penn House

Another hotel popular with some visitors is the Doubletree at 1515 Rhode Island Avenue Northwest which is near N st.



Here people can suggest vocabularies and other things to work on. Just put a topic in the list and other people can put their names behind it to indicate interest in that topic.

Topics suggested after June 3 morning sessions:

W3C Basic Geo work SWEET Overview Topography & Landforms Modeling ODPs Bottom-up Methodology / Clustering Triplifying GeoData POI/Placename Land Use Ontology Geometry Vocab Review/Southampton Vocabs OSM / LGD GeoNames/Gazetteer issues

Vocabularies, etc. I want to to work on[edit]

  • Core Vocabularies Discussion & Development (GeoSPARQL feature/geometry model, GeoSPARQL spatial relationships, feature types/points of interest, events/time/change)
  • Identify Auxiliary Vocabularies (Common Dataset/Named Graph geo metadata techniques, GeoProvenance, GeoVOID, additional spatial relationships(size,distance,direction,etc.)...)
  • Identify needed Open Source tools/extensions (GeoSPARQL, RDFizers, RDF publishing frameworks extensions, slippy map extensions, linked data browsers, etc.)
  • Task vocabularies to enable semantic description of geo workflows
  • Extending to example application vocabularies in applied fields (such as landcover)

How Much[edit]

The VoCamp event itself is free, although participants will need to pay for their own travel and accommodation.



  • Ed Summers
  • Gary Berg-Cross (SOCoP and INTEROP)
  • Todd Pehle (firstname_lastname(@)yahoo)


  1. Michael Lang (Revelytix)
  2. Krzysztof Janowicz (PSU)
  3. Gary Berg-Cross (SOCoP INTEROP & Knowledge Strategies)
  4. Mike Dean
  5. Marco Neumann
  6. Ed Summers (LoC)
  7. Fred Fonseca (PSU)
  8. George Thomas
  9. Gaurav Sinha (Ohio University)
  10. Ryan Kohl (Praxis Engineering)
  11. Nancy Wiegand (University of Wisconsin)
  12. Dalia Veranka (U.S. Geological Survey)
  13. Dave Mattli (U.S. Geological Survey)
  14. DEBORAH MACPHERSON CSI, AIA (First day only ) Cannon Design
  15. Mark Gahegan (Director, Centre for eResearch & School of Environment University of Auckland, New Zealand
  16. Dave Kolas (BBN)
  17. Ola Ahlqvist (Ohio State- remote or in person)
  18. John Moeller
  19. Naijun Zhou (UMD)
  20. Josh Lieberman (1st day)
  21. Todd Pehle (Orbis Technologies)
  22. Wei Luo (PSU)
  23. Sen Xu (PSU)
  24. Alexander Savelyev (PSU)
  25. Fabian Neuhaus (NIST, Friday only)
  26. Paulo Raposo (PSU)
  27. Rajashree Datta (Friday only Hunter College and BoK2 project)
  28. Rob Raskin (Jet Propulsion Laboratory, NASA)
  29. Dave Smith (USEPA)
  30. Jaymes Cloninger (US Army, SYNCADD)
  31. David Dettloff (US Army, SYNCADD, Friday only)
  32. Ruixin (Ray) Yang (Associate Professor, Geography and Geoinformation Science, GMU)
  33. Phillip Cousins (Sofi Executive Systems LLC)
  34. Robert Battle (Raytheon BBN Technologies, Friday only)
  35. Plamen Petrov (BBN)
  36. Luis Bermudez (OGC, 1st day)
  37. Leo Obrst (MITRE, Possibly on 2nd day)
  38. James W. Wilson (James Madison University -2nd day only)
  39. Steven Johnson (1st day)
  40. Dianne Davis (buildingSMART alliance)
  41. Bernadette Hyland (3 Round Stones Friday only)

Would like to, but can't[edit]

  1. Andreas Harth (but check out
  2. Juan Salas

Local Stuff[edit]

  • The Library of Congress Cafeteria in the Madison Building. Just take the elevators outside the vocamp meeting room to the 6th floor. They serve breakfast and lunch, and should have coffee. Not exactly fine dining, but it'll do in a pinch.
  • Pete's Diner just outside, across the street on the North side of the Madison Building. Great combination of typical diner food with a selection of asian style specials. Definitely try the Vegetarian Combo.
  • Firehook Bakery a good place to get a cup of coffee, muffin or a sandwich. If you prefer, Starbucks is half a block further down.

Outcomes & Topics[edit]

We will be adding content as topics for worksessions are decided.

Some ontologies can be seen the ontology community visible at:

You can join the community to add to and edit the ontologies there.

Likely topics for workshop discussion include:

  1. Linked Data Geo Vocabulary
  2. Ontologies for Linked Spatiotemporal Data
  3. USGS topics and
  4. Land Use and Landforms

Related Events and Calls[edit]