[OSGeodata] IRC Meeting this Thursday - 1500UTC

[Jo Walsh]

dear Stefan, all, 
On Thu, Aug 10, 2006 at 11:07:07AM +0200, Stefan F. Keller wrote:
> http://www.geometa.info/rappiinfo/wiki/index.php/OSGeodata, and moved the
> OAI-PMH vs. WFS comparison to
> http://www.geometa.info/rappiinfo/wiki/index.php/OAI-PMH

Thanks very much for this! (saved me a lot of effort, too.)
I put together a very simple and partial implementation of an OAI-PMH
interface. One thing that concerns me is that it's all about data
'provision' and not so much about 'production' - needs a WFS-T
equivalent. Right now it returns hideously ugly RSS by default. 
I agreed with your comment in earlier mail about needing a better
namespace than fgdc, fgdc-plus to express things with. Maybe 'geometa'
would be fine for this. 

> I think that two crucial questions are crucial:
> i) Do we need a search protocol or a harvesting protocol?
> ii) How to describe data access services as well as filter services?

I agree with both you and Norm about this. I think if we get the inner
model right, make it really useful to maintainers and producers, then
the discovery(search) / exchange(harvesting) side almost will 'drop out';
if we build something useful to us and use it for goodness, then
others are more likely to 'buy in' and there's a good chance of this. 

> Regarding ii) I would put the geographic data/resources in the focus being
> pointed to by a metadata entity. There we have to define enumeration values
> for data access services (like file api, jdbc, http get, wms, wfs) being a
> specialized dc-element 'type'. On the other hand we need to model filter
> services/resources as other metadata entities on their own (e.g. format
> conversion service or coord transformation service).

Right, that's interesting, INSPIRE talks about service translation
services and i was never sure what that was referring to - maybe they
just don't exist yet other than in the minds of eurocrats? ;)

cheers,


jo