[GeoSciML] Some observations on the CGI vocabs

Simon.Cox at csiro.au Simon.Cox at csiro.au
Mon Oct 22 21:32:28 EDT 2012


Steve -

Sorry - I spotted a couple more refinements that I think would assist in deployment:


1.       Make the ontology URI http://resource.geosciml.org/vocabulary/cgi/201202/alterationtype etc, and make the actual file name "alterationtype.rdf" etc. The Ontology URI is the identifier for the container resource (file) which can bring together things from all sorts of RDF namespaces, but should be gettable as a whole from the URI. There is a tradition of appending a # to the ontology URI as the default namespace, but this is not required and only applies if you are using # namespaces which generally don't work for vocabularies (as opposed to ontologies). Furthermore, this is an address where we can deploy the actual files and let Apache take care of dispatching. It gets the date from the directory.

2.       Make the concept scheme http://resource.geosciml.org/classifierscheme/cgi/201202/AlterationType etc, i.e. cgi:AlterationType etc. That fits the CGI URI policy, ties the date to the scheme, and is not the same as the ontology URI
Attached are revised examples.

Simon

From: Cox, Simon (CESRE, Kensington)
Sent: Monday, 22 October 2012 2:09 PM
To: Steve.richard at azgs.az.gov
Cc: geosciml at lists.opengeospatial.org
Subject: Some observations on the CGI vocabs

Hi Steve - prompted by the flurry of email today, I've taken a look at a couple of the vocabs.
Good work. I'm seeing some basically minor issues that would make them even better ...


1.       For each of the vocabularies, you have used the Ontology URI to identify the ConceptScheme. This is beyond OWL-DL (or OWL2-RL), so is generally frowned on by DL-nazis. Mind you, these days I'm relatively unconvinced that ConceptSchemes provide much utility, so I just use a placeholder URI, like AlterationType:Scheme.

2.       OTOH I find that a skos:Collection for each partial URI, ending in "/", helps navigation, as you can use skos:member to point to all the initial members.

3.       Not clear why all the concepts are type=owl:NamedIndividual as well as skos:Concept?  And is it necessary to include the superclass owl:Thing?

None of these will have much direct effect on the things that matter (the concepts, the concept hierarchy, labels and definitions) but will make the vocabularies more digestible to SemWeb types.

I've processed two examples where I have done the suggested remediations.

Simon







[cid:image001.jpg at 01CDB0FE.2CC656B0]

Simon Cox | Research Scientist
CSIRO Earth Science and Resource Engineering
ARRC, PO Box 1130, Bentley WA 6102, Australia
Tel +61 8 6436 8639 | Mob +61 403 302 672
simon.cox at csiro.au<mailto:simon.cox at csiro.au> | http://csiro.au/people/Simon.Cox<http://www.csiro.au/people/Simon.Cox>



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengeospatial.org/pipermail/geosciml/attachments/20121023/4c22eefe/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 5458 bytes
Desc: image001.jpg
URL: <http://lists.opengeospatial.org/pipermail/geosciml/attachments/20121023/4c22eefe/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: alterationtype.rdf
Type: application/octet-stream
Size: 33971 bytes
Desc: alterationtype.rdf
URL: <http://lists.opengeospatial.org/pipermail/geosciml/attachments/20121023/4c22eefe/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: faulttype.rdf
Type: application/octet-stream
Size: 43693 bytes
Desc: faulttype.rdf
URL: <http://lists.opengeospatial.org/pipermail/geosciml/attachments/20121023/4c22eefe/attachment-0001.obj>


More information about the GeoSciML mailing list