RE: RDF RSS sitemaps [Re: Survey of RDF data on the Web]

The RDDL spec [1] covers this kind of thing quite well, mostly using XLink*

Personally I'd very much like to see some kind of consensus on a 'preferred'
approach - I've got this very issue coming up myself in the near future, for
replacing an (already pretty good) XML-based filesystem indexing scheme**.

Cheers,
Danny.

[1] http://www.rddl.org

* don't know whether this'll be an issue, but XLink seems to be back in the
air again thanks to XHTML 2.0, see xml-dev

** JavaHelp (a Java extension package from Sun) uses 4 xml files - Map file
(one-to-one mapping between logical names and file locations), Index
file(associations between logical names & book-style index list entries),
TOC file (hierarchical table of contents showing sections, chapters, pages
etc, logical names structured using xml nesting) Helpset definition file
(points to the other 3 file).
Given these 4 files JavaHelp provides a neat navigation UI. More info at Sun
&
 http://www.isacat.net/2001/code/autohelp.htm




---
Danny Ayers
<stuff> http://www.isacat.net </stuff>

Idea maps for the Semantic Web
http://ideagraph.net


>-----Original Message-----
>From: www-rdf-interest-request@w3.org
>[mailto:www-rdf-interest-request@w3.org]On Behalf Of Seth Russell
>Sent: 20 August 2002 19:05
>To: Chris Croome; www-rdf-interest@w3.org
>Subject: Re: RDF RSS sitemaps [Re: Survey of RDF data on the Web]
>
>
>
>From: "Chris Croome" <chris@webarchitects.co.uk>
>
>> I've been thinking about this as well -- I generally put metadata for
>> web pages in an associated DC RDF file and link to it using <link
>> rel="meta" .../>, one RDF file per XHTML document.
>>
>> This is fine if you are happy finding all the RDF files via XHTML pages
>
>Which of course we are not.
>
>> but in order to enable access to all the RDF files directly I was
>> thinking that perhaps a RSS 1.0 sitemap listing all the RDF files for a
>> site might be a good thing.
>>
>> If RSS is used nothing new is needed, something like this would do:
>>
>>   http://mkdoc.com/rss100sitemap.rdf
>>
>> But with the items pointing to the RDF files and perhaps with the
>> addition of a <dc:format> for the mimetype.
>
>Good idea, this works for me too :)
>
>However it is a bit indirect for identifying a resource which is
>readable as
>a semantic web document:
>
>1) First you need to visit the node of the subject Resource
>2) Then you need to check that the value of dc:format is
>"application/rss+xml" or "application/rdf+xml".  Also alternate
>formats like
>N3 and Quads that don't have registered media types are left hanging.
>
>In my application that extra time to locate this potential
>information would
>probably be prohibitive.  I like to put a ~read~> button beside the link to
>any resource which I can read into my sailor agent (see screen
>shot).  Maybe
>we could use a new property ... call it semref .  We would use
>this property
>to point out a resource that we can read like RDF.  This proerty
>would be to
>semantic web documenta as <a  href> was to HTML documents.  Such a property
>would also provide a direct path for spiders looking for semantic web
>metadata files ... much more efficient and reliable than seeAlso.
>
>Seth Russell
>
>language: semenglish
>
>semref
>    type  Property;
>    domain  Resource;
>    range  Semdocs;
>    subPropertyOf  seeAlso;
>    comments  "Indicates a semdoc that is related to the subject resource."
>;
>    isDefinedBy [
>        semref <http://robustai.net/sailor/sailorOntology.quads>;
>        seeAlso (sailor agents)
>    ].
>
>(sailor agents)
>    seeAlso <http://robustai.net/sailor/>;
>    semref
>        <http://robustai.net/sailor/sailorOntology.quads>,
>        <http://robustai.net/sailor/semdocs.quads>;
>    screenShots
>        <http://robustai.net/sailor/screenshots/scuterplan.html>.
>
>
>

Received on Wednesday, 21 August 2002 05:35:48 UTC