RE: [topicmaps-comment] referring to a topic from outside a TM (w as:multilingual thesaurus - language, scope ,and topic naming constrain t)

["Bandholtz, Thomas" <thomas.bandholtz@koeln.sema.slb.com>]

Title: RE: [topicmaps-comment] referring to a topic from outside a TM (was: multilingual thesaurus - language, scope ,and topic naming constraint)

Lars,

I think I have already answered most of your comments in http://lists.oasis-open.org/archives/topicmaps-comment/200203/msg00037.html.

To summarize:
* if the URI reference does not really address something physically, it will only be a name (still an URI reference).
* the meaning of a fragment identifyer is only defined in relation to certain media types.
* this seams to be implemented for HTML only till now.
* a database service (i.e. using a Java Servlet) has no pre-detectable media type, so the Web server/Servlet engine will not pass the fragment identifyer at all.

* A *query* (indicated by an ? in an URI) is not a fragment identifyer, but part of the URI (not: URI reference) itself, so the meaning is not dependant of the media type.

* There seems to be no implementation of query-support for HTML files.

--> If a PSI shall be a persistant URI able to access a Topic definition physically - you must never change the "media type" (including data base) of this definition :-)

Another question is: what do you really access?
*** Fragment identifyers address sequential positions (points) in documents that are of random sequence. They do not address objects. What is behind your access point - up to the "end of file" - is completely undefined, and there is no way to define it.

*** The meaning of a query is completely undefined at all, but there is a way to define it. I am currently preparing a Web Service Description (WSDL) sample that exactly defines that an URI like http://your.domain.xyz/psi/services/getTopic?id=4711 will return exactly the definition of Topic 4711.  This WSDL is also including the schema that will be used in the response.

But - what shall be returned when you access an association? The association only, containing unresolved links to its topics? The association plus these topics? There are different "chunk" (as Thomas Passin said) options, and I am afraid we need a concept of a <topicMapFragment> to get this solved.

Thomas Bandholtz
XML Network
Competence Center Content Management
SchlumbergerSema
Sema GmbH
Kaltenbornweg 3
D50679 Köln/Cologne
+49 (0)221 8299 264
 

> -----Original Message-----
> From: Lars Marius Garshol [mailto:larsga@garshol.priv.no]
> Sent: Thursday, March 14, 2002 7:47 AM
> To: topicmaps-comments
> Subject: Re: [topicmaps-comment] referring to a topic from
> outside a TM
> (was: multilingual thesaurus - language, scope ,and topic naming
> constraint)
>
>
>
> * Thomas Bandholtz
> |
> | Try this with http://psi.ontopia.net/ontopia/ontopia.xtm#topicmap in
> | IE 6.0. You will see the top (!) of Steve Pepper's XTM-document,
> | while the topic with id="topicmap" resides two scrolls below. Is
> | this a bug or a feature?
>
> Depends what you are after. If you can explain why you want to resolve
> the URI we can start to unravel the implications of what happened. But
> the crucial issue is really: what were you trying to achieve when you
> resolved that URI?
>
> BTW, the last URI you came up with is different from the first two. It
> points to a different resource, and passes a parameter to it. What the
> parameter means (and so also what the entire URI means) is up to the
> application. So for you to expect to get a topic map from this URI is
> entirely unreasonable. You have no reason to expect such a thing.
>
> |  http://psi.ontopia.net/ontopia/ontopia?id=topicmap.
>
> --
> Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
> ISO SC34/WG3, OASIS GeoLang TC        <URL:
> http://www.garshol.priv.no >
>
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>
>