4. Hundreds of millions of existing occurrence records already live in relational databases with their own choice of geospatial and other vocabularies, some standard, some not. Except in the few(?) cases that their Relational schema is DwC, they have to map their local schemas to DwC for their service responses. It's likely inconsequentially more work to map to some other vocabulary that is also semantically similar to theirs than to map to DDwC. The main burden will fall on Markus and other tool builders who help make mapping issues invisible to undersupported data providers in ways that don't depend very much on what the mapping is from. They'll have robustness requirements imposed on them which become a matter of choice when the provider and the map maker are one in the same.
On Mon, Nov 22, 2010 at 10:48 AM, Peter DeVries pete.devries@gmail.com wrote:
Hi Markus,
In general I feel lots of the latest threads (though I have to admit I hadnt had the time to read them all) seem to rather discuss the (re)creation of a new tdwg ontology instead of suggesting changes to darwin core. Darwin Core is ratified and I think we should refrain from taking every single term it has to pieces, redefine its definition and reconstruct it in different ways because of linked data demands.
We might be better off to start something new for the LOD world.
I agree. For three main reasons,
- The current DarwinCore works for what it was originally intended.
- Creating something that works well for the semantic web is somewhat
tricky and will take some time. 3) I think this discussion is confusing potential DarwinCore users. My main goal was to figure out a "home" for this ontology, and it seems as if TDWG only wants to use things in it's namespace. It is my feeling that other groups like EUNIS will want to use this so I would like to figure out the final ontology uri soon. I could add this to the txn ontology, but I think it has utility even outside biodiversity informatics.
- Pete
Here is an example with some of the non-relevant stuff taken out.
<dwc_area:Area rdf:about="geo:44.86528100,-87.23147800;u=10"> dcterms:title44.86528100, -87.23147800 Radius 10 meters</dcterms:title>
dcterms:identifiergeo:44.86528100,-87.23147800;u=10</dcterms:identifier> dcterms:created2010-10-28T00:00:00-0500</dcterms:created> dcterms:modified2010-11-19T22:17:37-0600</dcterms:modified> geo:lat44.86528100</geo:lat> geo:long-87.23147800</geo:long> <dwc_area:radius>10</dwc_area:radius> <dwc_area:areaWithInFeature rdf:resource="http://sws.geonames.org/5250768/%22/%3E wdrs:describedBy rdf:resource="http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9.rdf"/ </dwc_area:Area>
What I have done with my examples is standardized on a fixed precision to the lat and log.
The level of precision does not have be this high, we just need to have an agreed on level of precision. (Probably something that we should do anyway since the actual precision should not be inferred by the significant digits.)
So Steve's 44.86,-87.23 => 44.86000000, -87.23000000.
Now we have a standard urn-like entity for an area that follows the proposed ietf standard.
Other data like soil type, weather etc can be tied to these areas.
Area1 => hasDegreeDayValue "2105"
Other areas can be related to this area Area1 within Area2, AreaB near AreaC etc.
AreaZ within NaturePreserveB
Also the following was recently proposed by the ietf.org
A Uniform Resource Identifier for Geographic Locations ('geo' URI) http://tools.ietf.org/html/rfc5870 (from Sean Gillies)
- Not clear that all the systems understand this but URIburner and
Virtuoso interprets these as a URN type thing so they work but are not understood in the way that the "geo" vocabulary is understood. Note that if and when this becomes a standard it will allow these "Areas" to be universally understood.
What this means is that I can markup my 10,0000 mosquito records from one location with one URN type id.
In summary, these allow more efficient occurrence records that leverage the existing support for the geo vocabulary and the probably support for the ietf.org standard.
I have added some predicate to the vocabulary that allow inferencing from the area through the geonames hierarchy.
This would allow one to infer the following about the record above.
Area > areaWithInFeature => Door County => Wisconsin => United States => North America = Earth
This does not mean that additions could also be made that allow inferencing up some other non-geonames hierarchy.
Here is the ontology http://lod.taxonconcept.org/ontology/dwc_area.owl Here is the owl doc http://lod.taxonconcept.org/ontology/dwc_area_doc/index.html
Here is an example of an occurrence record that uses this:
HTML http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9.html RDF http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9.rdf
Respectfully,
- Pete
On Fri, Nov 19, 2010 at 12:32 PM, Steve Baskauf steve.baskauf@vanderbilt.edu wrote: Pete, I was going to ask questions about this the first time you mentioned it, but got distracted. I guess the main question I have is: what you would "do" with it? I guess it could be considered an identifier for a spot on the earth, but based on what I've read about guids it's considered to be a "no-no" to try to infer stuff about a resource by looking at the form of the identifier. Rather one should look at the metadata associated with the identifier to understand things about the identifier (which you provide with geo:lat and geo:long). I suppose that one could consider this to be some kind of identifier that could be reused, but particularly since the precision of your lat and long are 8 digits, it is highly unlikely that anybody besides you is ever going to choose to use this identifier over (vs. 44.86,-87.23 which would be imprecise enough to include a lot of places to which people might want to refer). I may just be misunderstanding the purpose you intend.
The other question is a more general one. Do we need more ways to specify uncertainty in location than we already have? We already have dwc:coordinateUncertaintyInMeters and dwc:coordinatePrecision . I've been using dwc:coordinateUncertaintyInMeters with a seat-of-the pants estimate on my part about how accurate I think my geolocation is (expressed in meters). That may be a misuse of this term because I'm really thinking radius around a point rather than uncertainty of coordinates. But as a practical matter, if I think my estimate of location is good to 1000 m (vs. 100 m or 10000 m) does it really matter if I'm talking about a square or a circle? In any case, I'm saying, this lat/long could be off from the actual location by a km. If I had a GPS receiver that allowed me to download the actual estimated accuracy (based on satellite signals and whatever else), then I would populate dwc:coordinateUncertaintyInMeters with that, but mine crummy old one doesn't. To me the most important thing is for users to know whether this is a ballpark estimate or if they could expect to actually be able to walk up to the tree using the coordinates they give.
Steve
Peter DeVries wrote:
I wrote about this earlier but I never heard anything back.
I have made something that uses the geo vocabulary but also allows pointRadiusSpatial fit measure that I call radius.
The advantage is that this adds a standard way to deal use something like an extent or pointRadiusSpatial while still benefiting from the widely used geo vocabulary.
It also allows these "Areas" to be referenced in a commonly understood urn way that using a ietf standard.
For example: "geo:44.86528100,-87.23147800;u=10"
There are still some things I need to fix and check with this vocabulary but I am wondering if there is any interest in incorporating this into the DarwinCore.
If not I will probably change the name of the ontology.
There are also things in the example below that are not part of my proposal.
I have what I call "Areas" that look like this:
<dwc_area:Area rdf:about="geo:44.86528100,-87.23147800;u=10"> dcterms:title44.86528100, -87.23147800 Radius 10 meters</dcterms:title> dcterms:isPartOf rdf:resource="http://lod.taxonconcept.org/ontology/void#this"/
dcterms:identifiergeo:44.86528100,-87.23147800;u=10</dcterms:identifier> dcterms:created2010-10-28T00:00:00-0500</dcterms:created> dcterms:modified2010-11-09T16:33:34-0600</dcterms:modified> geo:lat44.86528100</geo:lat> geo:long-87.23147800</geo:long> <dwc_area:radius>10</dwc_area:radius> txn:elevation186.54</txn:elevation> txn:continentNorth America</txn:continent> txn:countryCodeUS</txn:countryCode> txn:countryUnited States</txn:country> txn:stateProvinceWisconsin</txn:stateProvince> txn:countyDoor</txn:county> txn:localityTextTown of Sevastopol</txn:localityText> txn:locationNameShivering Sands Natural Area Woods</txn:locationName> txn:areaHasOccurrence rdf:resource="http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9#Occurrence"/ txn:areaHasObservedSpeciesConcept rdf:resource="http://lod.taxonconcept.org/ses/ICmLC#Species"/ txn:areaHasIndividual rdf:resource="http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9#Individual"/ txn:areaInStateProvince rdf:resource="http://sws.geonames.org/5279468/"/ <txn:areaInCounty rdf:resource="http://sws.geonames.org/5250768/%22/%3E wdrs:describedBy rdf:resource="http://ocs.taxonconcept.org/ocs/f522444a-2dd9-400e-be59-47213ef38cb9.rdf"/ </dwc_area:Area>
I recently added the following predicates, but have not altered my RDF examples.
#featureContainsArea #areaWithInFeature
http://lod.taxonconcept.org/ontology/dwc_area.owl
OWL Doc http://lod.taxonconcept.org/ontology/dwc_area_doc/index.html
The predicates are a bit awkward, but I wanted to be clear that this was to link an "Area" like "geo:44.86528100,-87.23147800;u=10" to a Geonames "Feature".
I thought a different set of predicates could be created to deal with some other class of "SpatialThing" if needed.
Respectfully,
- Pete
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 TaxonConcept Knowledge Base / GeoSpecies Knowledge Base About the GeoSpecies Knowledge Base
-- Steven J. Baskauf, Ph.D., Senior Lecturer Vanderbilt University Dept. of Biological Sciences
postal mail address: VU Station B 351634 Nashville, TN 37235-1634, U.S.A.
delivery address: 2125 Stevenson Center 1161 21st Ave., S. Nashville, TN 37235
office: 2128 Stevenson Center phone: (615) 343-4582, fax: (615) 343-6707
http://bioimages.vanderbilt.edu
--
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 TaxonConcept Knowledge Base / GeoSpecies Knowledge Base About the GeoSpecies Knowledge Base
<area_sparql_query.jpg>
--
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 TaxonConcept Knowledge Base / GeoSpecies Knowledge Base About the GeoSpecies Knowledge Base
tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content