[cat-interop] GitHub and registry URI dereferencing

[Signell, Richard]

+1 on the GitHub idea.  Since we have no standard now, I would vote
for OSGeo for starters, where a group of folks can try implementing in
a few catalog services (e.g a few geonetwork and geoportal servers),
demonstrate how this improves interoperability by building a few
IPython Notebook examples using OWSLib, and then later we can propose
to OGC or ESIP for further discussion?

On Wed, Mar 5, 2014 at 11:44 AM, Steve Richard
<steve.richard at azgs.az.gov> wrote:
> I like the idea of having the registry on GitHub, using a neutral domain for
> the host name in the HTTP URIs, with a redirector to point those URIs at
> authoritative resources when dereferenced.  These resources must be under
> some stable stewardship, ideally not linked with any commercial interest,
> and with a supporting community. Should the host be metadata101  or
> OSGeo--which is more stable and better supported.
>
>
>
>
>
> Stephen M Richard
>
> Arizona Geological Survey
>
> 416 W. Congress  #100
>
> Tucson, AZ
>
> AZGS: 520-770-3500
>
> Office: 520-209-4127
>
> FAX: 520-770-3505
>
>
>
> From: Tom Kralidis [via OSGeo.org]
> [mailto:ml-node+s1560n5107713h84 at n6.nabble.com]
> Sent: Wednesday, March 05, 2014 9:31 AM
> To: Steve Richard
> Subject: Re: [cat-interop] Metadata link types vocabulary proposal
>
>
>
>
> Thinking this though more, I think the way forward here is for this
> to be an OSGeo vocabulary, which we can build and manage under
> http://osgeo.org/vocab like:
>
> http://osgeo.org/vocab/data/{provider}/{type}[/version]
> http://osgeo.org/vocab/service/{provider}/{type}[/version]
>
> Then providing a thin resolution layer, members of this vocabulary, when
> invoked via HTTP, would then dereference/resolve accordingly.  Like:
>
> http://osgeo.org/vocab/service/ogc/sos/1.0
>
> would resolve to:
>
> http://www.opengis.net/sos/1.0 (which then resolves as per the
> authoritative provider's wishes).
>
> This provides the following benefits:
>
> 1./ metadata clients/servers can bind to the OSGeo vocabulary, thus
>      realizing the core goal of a harmonized link type list
> 2./ anyone who really wants to invoke / resolve these can do so
> 3./ This keeps OSGeo as a just a gatekeeper / custodian of the list with
>      redirects as appropriate, which gets us out of the complexity of
>      managing these across standards orgs/etc.
>
> Comments?  I've written a small Flask app to act as a high level
> resolver, based on our link types list and will resolve / redirect
> accordingly, which we can push forward for hosting, etc.
>
>
>
> On Mon, 3 Mar 2014, Tom Kralidis wrote:
>
>
>> Date: Mon, 3 Mar 2014 21:53:51 -0500
>> From: Tom Kralidis <[hidden email]>
>> To: Paul van Genuchten <[hidden email]>
>> Cc: "<[hidden email]>" <[hidden email]>
>> Subject: Re: [cat-interop] Metadata link types vocabulary proposal
>>
>> On Mon, Mar 3, 2014 at 4:07 PM, Paul van Genuchten
>> <[hidden email]> wrote:
>>> I support Stephan at this point. The semantic web can be a big consumer
>>> of spatial data services, if we're able to optimally embed our data in
>>> it. Linking to controlled vocabularies by url is a big facilitator.
>>>
>>
>> Good points.
>>
>>> Sure i see the point of having a urn structure as
>>> {provider}:{type}:{identifier}:{version}. But this can also be managed
>>> in a (skos) vocabulary-identifier-as-url by using identifiers like:
>>>
>>> for example:
>>> http://osgeo.org/vocab/service/ogc/wms/1.1
>>> http://osgeo.org/vocab/service/opensearch/opensearch/2.0
>>>
>>> and
>>>
>>> http://osgeo.org/vocab/mediatype/esri/shapefile/1.0
>>> http://osgeo.org/vocab/mediatype/ogc/gml/2.0
>>> http://osgeo.org/vocab/mediatype/w3c/json-ld/1.0
>>>
>>> or
>>>
>>> http://osgeo.org/vocab/service#ogc_wms_1_1
>>> http://osgeo.org/vocab/mediatype#ogc_gml_2_0
>>>
>>> At this point we should consider if osgeo.org is the best place to set
>>> up such a service/vocab, as this is more a data-matter then a software
>>> matter. Should we connect with geovocab, w3c/ogc, dbpedia to have this
>>> set up?
>>>
>>
>> Based on the comments on HTTP URIs at
>> https://github.com/OSGeo/Cat-Interop/issues/3, the OGC has this at
>> www.opengis.net.  However the structure is not SKOS-like. I guess one
>> factor is how deep version negotiation would go.
>>
>> Then again, we could implement this in OSGeo and have resolver back to
>> their authoritative resources (i.e.
>> http://osgeo.org/vocab/service/csw/2.0.2 could resolve to
>> http://www.opengis.net/cat/csw/2.0.2.  Does this sound useful/make
>> sense?
>>
>>> If we feel osgeo.org is a good place, let's post about it in the
>>> request-list...
>>>
>>> _______________________________________________
>>> cat-interop mailing list
>>> [hidden email]
>>> http://lists.osgeo.org/cgi-bin/mailman/listinfo/cat-interop
>>
>
> _______________________________________________
> cat-interop mailing list
> [hidden email]
> http://lists.osgeo.org/cgi-bin/mailman/listinfo/cat-interop
>
> ________________________________
>
> If you reply to this email, your message will be added to the discussion
> below:
>
> http://osgeo-org.1560.x6.nabble.com/cat-interop-Metadata-link-types-vocabulary-proposal-tp5106675p5107713.html
>
> To unsubscribe from Cat-Interop, click here.
> NAML
>
>
> _______________________________________________
> cat-interop mailing list
> cat-interop at lists.osgeo.org
> http://lists.osgeo.org/cgi-bin/mailman/listinfo/cat-interop
>



-- 
Dr. Richard P. Signell   (508) 457-2229
USGS, 384 Woods Hole Rd.
Woods Hole, MA 02543-1598