Review & Revise with respect Catalog 2.0

Intiailal overview of Catalog 2.0 specification.

Metadata entity, and metadata element are still used extensively in the Catalog specificaiton. So readers of that specification will not find our classes entirely unexpected.

A full BNF is given for a query lanaguage (abilities are very similar to the Simple Feature for SQL).

CORBRA bindings are given, offering great insite into the expected mapping of Catalog Constructs into code. Unfortantly metadata is refered to as a "any"

HTTP / SOAP bindings are given - this may be the replacement of the WRS Dragt document from 1999?

Recomendations:

• Ape the CORBRA bindings as a starting point
• Do the usual java substitutions for take advantage of Collections, URL and so on
• Use MetadataEntity rather than "any"

Catalog 2.0 Specification also includes a great comparison of the ISO and OGC naming with respect to Metadata. A while ago we discussed finding the common ground - this document lists it.

Wild Idea:

• Implement a Adapter pattern for MetadataEntity allowing the support/mapping of multiple standards where useful.

Degree implementation is very good, represents a capture of ISO 19119 (Service Metdata) and ISO 19115 (Spatial Metadata) as beans (including construction from XML Elements).

Implementation of Catalog 2.0 constructs is faithful, although query is limited to the top level metadata elements. Thus separate query constructs exist for ISO 19119 and ISO 19115.

Recomend the following implementation:

• use degree code as example for naming (it is the same as the specification and it would prevent extra work on their part)
• only define accessors (for Catalog we are only concerned with Query not modification)
• extend MetadataEntity to allow a general purpose Query to be written

From Markus Muller's email:

To which WRS do you refer directly? The one defined by the public
discussion OGC paper 01-024r1? Then the answer is yes - and no .
Catalog 2.0 includes an http binding (Catalog Service for the Web).

This CSW is unfortunately abstract, or lets better say it is not sufficient for a "full" catalog. It is expected that profiles of the spec define the full interface and information model. There are currently two draft profiles. One defines a simple interfaces using ISO 19115/19119/19139 metadata. The other one is based on an ebRIM approach and defines an abstract meta-model for all kinds of registry objects.

CS-W as defined by Catalog 2.0 can nonetheless be implemented directly
The OGC is currently developing a test suite for it and a reference
implementation (deegree) that only covers this basic functionaliy. The
information model is based on Dublin Core.

Thanks you! This was the magic bit of information nobody was telling me. Be searching
for Stateless Catalog Profile I finally be able to find what happened to WRS.

Google found the following hit:
"OGC-IP- 2000-10-02 PAGE 1 TITLE: OGC Web Services Stateless Catalog Profile (was
Web Registry Service) VERSION: 0.05 EDITOR: Name: Lou Reich Address: Computer"

It really sounds like this is supposed to be the sequel:
http://www.intl-interfaces.com/servicemodel/StSCDiscussionPaperv3.pdf

A bit more searching found me this:

http://www.geoconnections.org/architecture/technical/specifications/geodata_registry/geodata_discovery_service_stateless.doc

Search marked it as version: 0.06 although the document says it is version 0.07:

The related document:
http://www.geoconnections.org/architecture/technical/specifications/type_schema_registry/type_schema_registry_service.doc

Seems to describe registering FeatureType.

Since catalog interfaces are not going to be included in next release, uncommit from 1.1. The work will be delayed to a future release.

The catalog work was considered pending and was not included in the 2.1 release.