[esip-semanticweb] ESIP namespace definitions

[Benno Blumenthal]

Not to gum up the works, but a bit of information that may be useful.

As it turns out, much to the surprise of many of us,  there is a problem
with converting (namespace,qname) pairs into URIs -- XML explicitly does not
define such a mapping (only the pair has meaning), and different XML
use-groups do it differently.   In particularly, RDF and RDF/OWL use simple
concatenation, while XML Schema inserts a # sign in between.   So

1) RDF style namespaces usually end in a # sign or at least a delimiter,
2) XML Schema style name spaces end in an ordinary character.


This also means that one has to choose between URIs being universal, and
(namespace,qnames) pairs being universal. In other words, if the URI is
universal (which it is supposed to be), then the namespace is not.



Benno




On Mon, Oct 26, 2009 at 3:19 PM, James Gallagher <jgallagher at opendap.org>wrote:

>
> On Oct 26, 2009, at 8:48 AM, Christopher Lynnes wrote:
>
>  We are now beginning to actually implement cross-ESIP applications that
>> have need of namespace declarations, specfically:
>> o  Federated search
>> o  Machine tagging of WMS capabilities documents
>> I believe there are some semantic web applications that are not too far
>> behind...
>>
>> Given that permanence is a desirable feature of namespace declarations, I
>> would like to propose we settle on a definition scheme that can be
>> implemented soon, but which we expect to be stable.
>>
>> I would like to suggest we use a scheme like:
>> http://esipfed.org/ns/foo/bar
>>
>> where
>> o  http is the scheme
>> o  esipfed.org is the authority
>> o  ns indicates that it is a namespace
>> o  foo/bar is a hierarchical path
>>
>> Examples of foo might be:
>> o  fedsearch
>> o  scast (for servicecasting)
>> o  datatypes
>>
>> So for example, to uniquely identify a URI as refering to a URL to a
>> browse image (useful for typing in Federated Search) responses), we might
>> have:
>>
>> http://esipfed.org/ns/fedsearch/browseImage
>>
>> One minor downside to this scheme is that it is often useful, but not
>> required, to put an informational document, such as a schema document, at
>> the URI.  Brian, would this be feasible?  Again, I don't believe it is
>> required for a namespace URI.
>>
>> Any counter-proposals?
>>
>
> A suggestion: That you include in the scheme a version number and that you
> do it like this:
>
> http://esipfed.org/ns/foo/bar/1.0#
>
> For example, the DAP 3.3 schema has the following in the root element:
>
> xmlns:dap3.3="http://xml.opendap.org/ns/DAP/3.3#" xmlns="
> http://xml.opendap.org/ns/DAP/3.3#"
>
> I think the version number is obvious - even HTTP has more than one
> version. Benno and Nathan requested the trailing '#' and as far as I know it
> exists to simplify something when building RDF although I've forgotten what.
> Regardless, it follows the Specs and also the OASIS recommendations. Maybe
> someone on this list can remind me regarding the trailing hash...
>
> Summary: Add a version number and I forget why we have the '#', but it
> might (will?) make working with RDF easier.
>
> James
>
> PS. Did you know that Hyrax now produces RDF? In the default configuration
> it's not enabled but you can see it in action at test.opendap.org (
> http://test.opendap.org/dap/data/hdf/contents.html and follow the 'rdf'
> links; e.g., http://test.opendap.org/dap/data/hdf/S2000415.HDF.rdf). It's
> still in testing --> really big files break it in a nearly deterministic way
> ;-)
>
>
>  --
>> Christopher Lynnes             NASA/GSFC, Code 610.2         301-614-5185
>>
>> _______________________________________________
>> esip-semanticweb mailing list
>> esip-semanticweb at lists.esipfed.org
>> http://www.lists.esipfed.org/mailman/listinfo/esip-semanticweb
>>
>
> --
> James Gallagher
> jgallagher at opendap.org
> 406.723.8663
>
>
> _______________________________________________
> esip-semanticweb mailing list
> esip-semanticweb at lists.esipfed.org
> http://www.lists.esipfed.org/mailman/listinfo/esip-semanticweb
>



-- 
Dr. M. Benno Blumenthal          benno at iri.columbia.edu
International Research Institute for climate and society
The Earth Institute at Columbia University
Lamont Campus, Palisades NY 10964-8000   (845) 680-4450
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.esipfed.org/pipermail/esip-semanticweb/attachments/20091027/7df33c3c/attachment.htm>