ASDI-L Mailing List Archive
From: David Crossley ([email protected])
Date: Wed Jun 18 2003 - 08:03:46 EST
Jonathan Doig wrote:
> Hi David
> Thanks for posting this work, which clarifies for me how the ISO 19115
> online resource information fields could work.
Great. Yes, demonstrations (even if they are not the final outcome)
sure do help to spark some ideas.
I am not too sure that it does "clarify". As i wrote in the technical
notes accompanying the demo and as Evert raised in his recent email,
the purpose of the ISO online resource fields are not well-defined and
so very obscure.
One thing that our demo does do, is to expose some of the issues.
The only real way to develop standards is by trying to implement them.
> We're undergoing a thorough re-design of the CANRI site to better
> expose and integrate our various websites, online map layers and
> metadata records for the user. In our current draft design, the user
> may browse or search by topic and specify a region, to see a
> Google-like results page, giving title, abstract and links to all
> three resource types as available. From there they can go to the site,
> add the layer to their map or view the metadata description.
So you have a separate metadata record for each layer. That is the
conclusion that i reached too.
> To support this "find-bind" use case, we'll need to revamp our
> registry of metadata about services and datasets (and it will also
> need to support the "publish" use case).
In the ASDD context i understand the "publish" case - describe something
and make it available at one of the nodes.
The "find" case requires a specific search to locate resources that
satisfy a certain need. In the case of our basic demo, we also search
for some additional parameters to restrict to services.
The "bind" case needs to make specific information available to an
application. In our demo, the xml response for the "Resource" provides
all of the linkage elements found in the resource description. Is that
sufficient to enable an application to use the resource? For Web Mapping
layers it is sufficient because all the information is encoded in the
> But I think we need to go beyond the model in your demo, where online
> services are just ways to get at the dataset. An online service may
> provide a subset of the underlying data with its own distinct
> attribute set and spatial footprint.
> So I think we have to treat services are resources in themselves, so
> the user can discover a service based on its own topic description and
> other criteria, distinct from the description of the underlying
> Eg the River Operational Data measures a bunch of water quality
> parameters, and one useful service from it just includes those sites
> where salinity is measured. This is really just a special case of the
> general problem of metadata for derived datasets, and more generally
> how you relate metadata records together to store and manage the
> information about datasets efficiently while still supporting
> discovery and description properly.
I would see that as a separate dataset, complete with its own metadata
The key question is, can we describe all "resources" within the ASDD and
provide sufficient metadata to discover and use them? I am not sure
about the answer yet.
Thanks for you excellent reply. It draws out some of the myriad issues.
> Jonathan Doig
> Jonathan Doig
> Program Director
> Community Access to Natural Resources Information
> Manager Natural Resource Information Coordination
> Department of Infrastructure Planning and Natural Resources
> Ph: 02 9895 7781
> Mobile: 0409 049185
> Fax: 02 9895 7834
> PO Box 3720 Parramatta NSW 2124