On Aug 10, 2005, at 9:15 AM, Matthew Cockerill wrote:
> This in fact ties back to my original question, when joining the
> list, about the best ontology(ies) to use for bibliographic data.
>
> Right now, BMC is planning to go the route of RSS+dublin core+PRISM.
> In order to ensure the interoperability of different bibliographic
> ontologies within piggy bank, my understanding is that this will
> depend on:
>
> (a) extending piggybank to support inferencing
> and
Correct. RDF provides a common model for integrating data. RDF Schema
and OWL build on RDF and (among various things) provide an
incremental set of relationships that license certain inferences.
This ties back to some of the previous 'owl tiny' discussions with
Jeen, et.al. from the Sesame team.
> (b) the provision of some glue RDF that expresses the various
> equivalence relationships between the different standards for
> bibliographic RDF markup.
Correct. But note there won't be one single bit of glue that works
for everyone.
> In terms of (b), my question is, how/where should reponsibility for
> that glue RDF sit, in order to make the naive users experience of
> PiggyBank as simple as possible. Ideally a new user should not have
> to navigate the web to find this RDF to teach their piggybank about
> bibliographic information.
>
> Could this information be maintained in a default, persistent,
> semantic bank, to which all piggy bank clients connect to by
> default, and which would provide a certain amount of curated
> 'background knowledge' that would help piggy bank to 'just work'
> from the point of view of end users.
Exactly.
What this puts in place is an ecosystem where different 'Banks' might
provide different services or "returns on your (data) investments".
(not sure of the Bank analogy works after all). You could imagine
some might even doing the heavier inferencing you mention in (a) and
simply make this additional information available.
But this again will be a matter of trust, convenience, subscriptions,
etc. In a corporate EDI environment, employees might subscribe to an
internal glue service. As I do scholarly research in a particular
environment I may choose my library and/or publisher banks. For the
web in general possibly useful 3rd party 'dataglue.com' services. I
suspect to even have my own personal glue I use to make sense out the
data that interests me. I'm not sure you'd want to subscribe to my
glue, but who knows. The design / model certainly would allow this.
A problem with the Piggy-bank / Semantic-bank scenario now is the
feedback loop. One can publish data from piggy-bank to a semantic-
bank but the notification back of new data in the collaborative space
"touches" data in my personal space is not there yet. I suspect once
this is in place, this will have interesting implications for gluing
up terms.
> [Of course a more sophisticated user might choose to exclude this
> default semantic bank, and instead define precisely what semantic
> information they wanted piggy bank to use...]
Yep :)
--
eric miller http://www.w3.org/people/em/
semantic web activity lead http://www.w3.org/2001/sw/
w3c world wide web consortium http://www.w3.org/
Received on Wed Aug 10 2005 - 14:28:31 EDT