[auscope-geosciml] WFS query with xlink:href [SEC=UNCLASSIFIED]
Oliver.Raymond at ga.gov.au
Oliver.Raymond at ga.gov.au
Sun May 15 19:13:16 EDT 2011
>> ACTION Marcus to check if Snowflake will allow us to query attribute
>> values in Xpath including xlink:href. (Result: tested during meeting
>> and it does.) ACTION CP to check if XtraServer will (tested after
>> meeting and it does).
A question for the Geoserver gurus. Does Geoserver have any problems with performing queries on attribute values in xlink:hrefs?
From: Sen, Marcus A. [mailto:mase at bgs.ac.uk]
Sent: Wednesday, 11 May 2011 7:10 PM
To: Raymond Oliver; Bruce.Simons at dpi.vic.gov.au
Cc: eboisver at nrcan.gc.ca; Guillaume.Duclaux at csiro.au; Simon.Cox at csiro.au; steve.richard at azgs.az.gov; portele at interactive-instruments.de; Duffy, Tim
Subject: RE: Testbed 4 - Geochemistry
> -----Original Message-----
> From: Oliver.Raymond at ga.gov.au [mailto:Oliver.Raymond at ga.gov.au]
> Sent: 11 May 2011 03:23
> I don't see a requirement for a GeologicSamplingMethod WFS.
> GeologicSamplingMethod is really just a pointer to a vocabulary, and we
> are adopting a best practice of using xlink:href and xlink:title to
> deliver vocabulary terms in GeoSciML.
We have a bit of an outstanding issue with our use of xlink:href's for remote property values. When we had a meeting with Clemens Portele in Edinburgh earlier this year he expressed a concern about some of our use of xlink:href where we just use the values of the URI or we are going to point to a vocab in RDF/SKOS. This is a bit of a mismatch with the GML feature model. We might keep some kind of GML serialization as an alternative to possibly be at the end of these href's as well as RDF/SKOS (or nothing where we just use the URI as an identifying value) so that Filter Encoding Spec queries can be done on their properties. Or we could define some new rules for how we handle encoding and querying non-GML property values in a GML Application Schema for putting to OGC. I guess we have implicitly settled on a best practice for our own use but should we document this explicitly for submitting some kind of recommendation to OGC?
I quote the relevant parts of our meeting notes below that Tim circulated to the auscope list on 7th March this year.
How to query by reference properties especially controlled vocab properties which we are now specifying with xlink:href URIs given that strictly we can't necessarily query on XML attribute values (they aren't properties in the GML feature model) and that their serialisation (if any) may be variable. E.g. RDF/XML won't nescessarily have a consistent XML encoding for same RDF model so an Xpath query isn't appropriate. Perhaps we need some kind of OGC change request for querying on identifiers which are URIs. (Also, general question of combining linked data triple type queries with GML feature model Xpath type queries.) ACTION probably up to wider community to work out querying XML and RDF data together.
We do need some mechanism for our queries on the identifiers of dictionary terms (mostly) that are URIs in the values of xlink:href attributes. This can be made to work by allowing attribute specifiers to appear in your query Xpaths but this is strictly outside the subset of Xpath that is required by WFS/FES and also is really abusing the intended role of xlink:href in GML which is really for pointing at a further serialisation and could be resolved by a WFS2 value-of() function so that the contents could be queried. In a pure GML implementation the identifier query would use a path name like .../value-of(gsml:geologicUnitType)/gml:identifier but we are now serialising our dictionaries in SKOS/RDF not GML. This is difficult in any case as it would require remote resolving capability, which is problematic in general for performance reasons and typically not supported be WFS implementations. CP will think about. Maybe we could add GML representations as well as SKOS that could be retrieved from the HTTP-URIs. ACTION Marcus to check if Snowflake will allow us to query attribute values in Xpath including xlink:href. (Result: tested during meeting and it does.) ACTION CP to check if XtraServer will (tested after meeting and it does).
Shouldn't this discussion be on the auscope list?
I've added Clemens to the CC list but when I used this address a yesterday I got an undeliverable response. Does anyone know if his email has changed or was this just a temporary problem?
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.
More information about the GeoSciML