[topicmapmail] How to address Web resources with XTM?
Jan Algermissen
algermissen@acm.org
Wed, 17 Mar 2004 12:01:11 +0100
Graham Moore wrote:
>
> Hi Jan,
>
> I am working on Semantic descriptions of REST web services using RDF and
> Topic Maps.
Yeah, exactly what I want to do. (Didn't know you were in the REST-camp...
Great!)
>
> So far I have adopted a pattern where the abstract notion of the service
> exists as a subject and thus is captured within the TM as a Topic with a
> subject indicator.
If I understand this correctly, the subject indicator does not use the
actual service URI but some sort of PSI, yes?
>
> Then because a given service or service capability could be deployed in
> different locations then I use either typed occurrence (end-point) or an
> association (again end-point to a topic with a subject address.
Ah, so you opt for using <resourceRef> to address the service, yes?
(My problem with this is that <resourceRef> addresses the bits and bytes
of a GET on the URI (at least in Web-land) and how can it then address the
service?)
>
> I think this a) preserves the subtleties of what is happening and b) is
> practical and useful for meta level services that utilise the descriptions.
What are 'meta level services'?
>
> In terms of resources I think this same idea can apply (and has done for a
> while in information science) where you make a distinction between the
> concept of the work and a representation of the work.
Hmm...I think, yes, sort of...you mean that you compare the work-edition
relationship to the software-deployedSoftware (Component-Artifact in UML2)
relationship, is that right. Interesting...
>
> I would like to get around to adding this pattern (or something like it) to
> topicmap central.
What do you mean by 'adding it to topic map central'?
>
> Do you think this is a helpful distinction to make?
I am not sure. Mu problem is that I am somehow in both camps now (RDF and
Topic Maps). I have long objected the complete REST-notion of resource because
of the (as I thought then) ambiguity of what URIs reference. At some point I
completely "got it" and I am now persuaded that there is actually no ambuguity
at all of what URIs identify....
My problem is that I now need to get this insight and Topic Maps aligned in
my head. And part of this is that Topic Maps only provide two different
addressing contexts where they should IMHO provide three (abstract concept,
service and information resource).[1]
Thus, I am maybe the wrong person to ask for help ;-)
Confusedly yours,
Jan
[1] No intention to propose any changes to Topic Maps, just trying to make
sense of all this in my head.
>
> - graham
>
> -----Original Message-----
> From: topicmapmail-admin@infoloom.com
> [mailto:topicmapmail-admin@infoloom.com] On Behalf Of Jan Algermissen
> Sent: 14 March 2004 12:56
> To: Lars Marius Garshol
> Cc: topicmapmail@infoloom.com
> Subject: Re: [topicmapmail] How to address Web resources with XTM?
>
> Hi Lars--
>
> (sorry for never following up on this thread..I am just too busy).
>
> Lars Marius Garshol wrote:
> >
> > * Jan Algermissen
> > |
> > | Because <resourceRef> is intended (IMHO) to address data (bits and
> > | bytes) and not abstractions of data. On the Web, topic maps
> > | neccessarily ignore the fact that a URL never refers to bits and
> > | bytes but the intention of <resourceRef> does not change by that.
> >
> > We agree on that. What I am saying is that the thing addressed by the
> > URI (the bits and bytes, as you say) is a different subject from the
> > abstract service itself, but that there is a relationship between
> > them. This I think you agree with.
> >
> > What I am saying is that this means you can create one topic for each
> > of the two subjects and represent the relationship between them as an
> > association.
> >
> > That should do what you want, shouldn't it?
>
> IMHO no. The problem I have is that we ultimately only have two
> possibilities (and I am not proposing to change that, BTW) to use a URI to
> address a subject:
>
> 1) via <resourceRef>
> 2) via <subjectIndicatorRef>
>
> In 1) we agree that the actual bits and bytes (the return of a GET) is the
> subject (ignoring the addressing indirection of HTTP)
>
> In 2) the subject is what the bits and bytes *are about* and these are
> usually not about the service itself that is producing the bytes.
>
> Furthermore, what if you have a service that does not even produce any
> representation (maybe only accepts POST and does processing on that data)?
>
> Again: I want to address the service and not what any bits and bytes
> obtained via GET are talking about.
>
> With RDF this is straightforward (just use the URI of the service in your
> statements) and I don't want Topic Maps to 'fall behind' here.
>
> Jan
> >
> > --
> > Lars Marius Garshol, Ontopian <URL: http://www.ontopia.net >
> > GSM: +47 98 21 55 50 <URL: http://www.garshol.priv.no >
> >
> > _______________________________________________
> > topicmapmail mailing list
> > topicmapmail@infoloom.com
> > http://www.infoloom.com/mailman/listinfo/topicmapmail
>
> --
> Jan Algermissen http://www.topicmapping.com
> Consultant & Programmer http://www.gooseworks.org
> _______________________________________________
> topicmapmail mailing list
> topicmapmail@infoloom.com
> http://www.infoloom.com/mailman/listinfo/topicmapmail
--
Jan Algermissen http://www.topicmapping.com
Consultant & Programmer http://www.gooseworks.org