[OSGeodata] geodata schema

[sfkeller at hsr.ch]

Norman,

You wrote on Tue, 1 Aug 2006 16:03:31 +0100: 
> I would be interested in taking the schema and the mind map that you
> have on the wiki and at
> https://geodata.osgeo.org/source/browse/geodata/trunk/metadata/ and
> developing it further.
> (the svn checkout for 'guest' seems to be broken - is there no guest
> access?)

Didn't realize this. This seems to borrow the information model from
FGDC. How much was this model discussed? For future compatibility, I
would either take an 'enhanced' set of attributes from pure core ISO
19115/19 - or an extended and specialized set of Dublin Core elements in
RDF, which would give geodata an even broader outreach through Google
with the help of OAI-PMH (c.f. below). Regarding FGDC and ISO 19115: I
found that the so called core lacks mandatory Dublin Core elements which
is a pity. 

> Although I know python quite well, I would prefer to develop in Java
if
> possible.
> 
> My motivation is to avoid the CSW-ebRim profile, and to encourage the
> uptake of RESTful catalogs, and further explore this rapid catalog
> development and see if enough metadata can be captured. My motivations
> are also from a business perspective - more catalogs mean more chance
to
> utilise meaningful data and a better use of clients.

That was exactly my intention too when I startet the thread "Ready for a
lean and mean Catalogue Service Protocol Spec.?" about ten days ago,
though perhaps more from a GDI and practical research perspective. 

> I would like to suggest the following work items in Java
> 
> (1) Take existing metadata postgis database
> (2) Use Java annotations (EJB3) to do a data-object persistence bind
> (3) Write a Java Bot to scour the web for getcapabilities and attempt
to
> parse the capabilities and fill up the database using (2) (want to
> abstract the database to set of POJOs)
> (4) implement georss feed (still thinking about this at the moment)
for
> new data items
> 
> Whilst I think 1 - 3 are quite easy, have there been any plans for
> develop the API for the catalog front end?
[...]

I like you plans (we are using Java too) and I support your approach,
"first implement then think" (I'm citing Jo here). Nevertheless just a
few thoughts here: What is needed is a protocol and an model/scheme for
the metadata. Regarding the metadata scheme I can live with both
approaches mentioned above given an incremental, implementation driven
process.

Regarding the protocol I see three choices: 1. CSW 2, 2. WFS and 3.
OAI-PMH. Though some say that CSW 2 (1.) can be profiled to a RESTful
protocol I simply don't see a reason why to deviate from WFS. This will
duplicate specification and implementation efforts. And both, 1. and 2.,
are somehow isolated in GIS world. Whereas 3. OAI-PMH is RESTful,
supported by Google and has many implementations which can be adapted. 

OAICat is one of the most active open source implementations I found:
http://www.oclc.org/research/software/oai/cat.htm. So, to repeat my last
question from 29 Jul 2006, was: Why no go with OAI-PMH (protocol)? 

Just my few cents here.