- From: Leslie Daigle <leslie@beethoven.bunyip.com>
- Date: Wed, 29 Mar 95 18:41:45 -0500
- To: hoymand@gate.net, ura-bunyip@bunyip.com, uri@bunyip.com
[hoymand@gate.net (Dirk Herr-Hoyman) wrote:]
> I'd like to know how you see URAs fitting into object brokers. While not
> claiming to have a full understanding of object brokers, they are meant to
> be independent entities on ANOTHER system (possibly) that goes out and
> finds the object in question. I believe the URN/URC resolution service
> that we have been batting around lo these many months would be an object
> broker here, and no doubt there could be others. It would seem, based on
> the discussions I see elsewhere, that object brokers should be part of your
> model.
With the caveat that I'm not completely versed in object broker proposals
either, let me plunge in with some replies :-)
At the level of models, I would see URAs fitting in with object
brokers in 2 possible ways -- by encapsulating information to contact
information brokers, or potentially as mechanisms for implementing
brokers. The former, of course, assumes that object broker systems will
be accessible through URI-conforming mechanisms.
> I would also like to know how you see URAs working with emerging
> object-oriented compound document specs such as OpenDoc. Again, my
> understanding is imcomplete on OpenDoc and its kin, but I believe that
> these too will have object brokers as part of their model.
URA's, as we've seen them, are not intimately tied to a particular document
or document type. Thus, they could operate on different types of document,
or from within different types of document. That is, you might build
an inter/active document with the URA's we are proposing, but that would
only be on application of the technology, and not necessarily its best.
In general, many of the document-related agents systems are meant to be
imbedded within documents, which tackles an importantly different problem from
the one we are proposing to address with URAs. That is, since URAs exist
independently of Web documents, they act on Web pages much as they might _any
other Internet information resource_, and are data objects in their own right.
This last point means that they can be separately stored, transmitted,
searched, and otherwise manipulated by (human and program) users.
Cheers,
Leslie.
------------------------------------------------------------------------------
"Been there, Leslie Daigle
Done that, leslie@bunyip.com
Read the newsgroup." -- ThinkingCat Montreal, Canada
------------------------------------------------------------------------------
Received on Wednesday, 29 March 1995 18:41:52 UTC