OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

tm-pubsubj-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: [tm-pubsubj-comment] Re: [topicmaps-comment] referring to a topic fromoutside a TM

Holger Rath
>> ... "Topic Map Addressing Scheme (TOMAS)"  ... could be based on a special URN scheme.
>> We know that TOMAS is *really* necessary and that it is not a  trivial
>> thing. But we will address it.

Thomas Bandholtz:
> I would not think about a special URN  scheme first. URNs are only names
> (ok, globally uniq names, hopefully). We need an addressing method that can
> provide on-line access to the addressed node.

Agreed. That's the first step. Then you have to figure what you retrieve when you get that
handle and pull it. Another story.

> Today most of us use URL with *fragment identifiers* (which is not widely
> implemented for XML) and not *queries* (using ? in an URI), so they are tied
> to documents.

You make an important point there. I think you made it already in other posts, but I
really catch it only now :))
As was said once in a debate with Steve N. and others, topic maps people have a tendency
to be "text bigots".
We boil all down to documents. And for that matter, we are certainly wrong.

XTM is a document interchange format, and certainly not fit for query. Ditto for any XML
format ...
BTW in Mondeca we use Oracle + Graph Manager. XTM is for import-export only, as could be
RDF or customized XML.

> "global knowledge interchange" will not be implemented by merging files or
> using file-based linking methods.
> I can only imagine a future for TM when they can be maintained in databases
> without needing a physical file representation. Linking and merging should
> just be "virtual" by processing the relations dynamically, using web interfaces.

Exact. Processing Gigabytes of XTM is crazy.

> The PSI activities today are dominated by file-oriented thinking.
> I think this is very risky.

I think you're right. In present recommendation draft of PubSubj TC
... there is no explicit reference in Part 5 to any form but "addressable documents" -
including XTM, RDF, XHTML.
But the notion of having PSI living in databases have been discussed in the TC, and I
think we should plough that notion further on, and include it in recommended practices.

OTOH look closely to the general recommendation - Part 4 - you will not find anything
contradictory with data base PSIs.
There has even been progress towards less document-oriented approach, since we've come
back to "Subject Indicator" after a passage through "Subject Definition Document". But we
have still "Published Subject Documentation Set", and the recommendation is about
"Documentation" of Published Subjects. Hmmm ...

Bottom line - Thomas, you are bringing interesting insights there. What about joining
PubSubj TC - as your Schlumberger colleague Mary Nishikawa?


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC