Matthew Cockerill wrote:
> Given that I think we agree an importent goal has to be to give end users something genuinely useful, that doesn't require to much technical expertise to comprehend, I guess that there's a need to decide which is most achievable:
>
> (a) getting websites, and scraper writers, to agree on what ontologies to use for important types of content, such as bibliographic content, and then focusing on tools that will work with that specific ontology
> or
> (b) encourage a free for all in terms of what ontologies to use, and then for any given domain: e.g. personal information, or bibliographic info, or locational info), set up a framework to allow the creation of "glue ontologies", that will describe the mapping between various most popular ontologies within a particular domain. (I'm making the terminology up here, as my understanding of the state of the art here is pretty limited.
>
> or
> (c) some balanced, pragmatic combination of (a) and (b).
(c) sounds like the best solution to me. The more people use the same
language the better, but we shouldn't be forcing them to do it.
I personally fell uncomfortable when forced to use somebody else's
markup... but given a choice, I would much rather reuse their than
invent my own :-)
A little child psychology at play, I guess.
> When you say: "Piggy Bank isn't supporting inferencing now", but that one day it might - is is structured such that there are appropriate hooks to allow it to be extended in that way. i.e. in the same way that screen scraper modules can be slotted in, can inference/mapping modules be dropped in. Is sufficient scaffolding in place to make a project to do bibliographic processing/inference feasible?
Equivalences, for example, are RDF statements, just like any other.
Some statements are 'passive' means that our system doesn't know how to
'react' on them, some others will be more active, means that their
presence will trigger particular actions on the data (for example, the
creation of new statements)
So, if you want to use your own ontology but you ship it with enough
mappings, then for us, once inferencing is available, it's totally the
same whether you used the original ontology or you used your own and
gave us a way to map it to those our tools understand.
Up to you.
--
Stefano Mazzocchi
Research Scientist Digital Libraries Research Group
Massachusetts Institute of Technology location: E25-131C
77 Massachusetts Ave telephone: +1 (617) 253-1096
Cambridge, MA 02139-4307 email: stefanom at mit . edu
-------------------------------------------------------------------
Received on Fri Jul 15 2005 - 19:32:54 EDT