[FoRK] base is out...
mattj at newsblip.com
Thu Nov 17 11:07:39 PST 2005
Quoting Luis Villa <luis.villa at gmail.com>:
> The interesting task for the next
> generation of metadata/semantic-web apps is how to make creating such
> structured data as easy and low-effort as possible.
That's actually what I see in Base's potential. It's not just
that will evolve in it; names for properties (name-value pairs) will evolve,
too. I think it's in the FAQ that user-defined property names that become
popular for a category will be presented as defaults. That means that a
category might start with naive property names like "Item", but can evolve to
use names like "SAPItemNumber". It means a dynamic semantics, rather than a
With transactional domains, where money is involved, that may be too
chicken-and-egg to start easily. But in the informational domain ("find me
movies playing near me similar to the movies I've rented this year"), where
rough results are tolerated, it might work out well. Worth watching,
> As to the response to this email... I think flickr and others are
> proving that relatively unstructured metadata can still be very useful
> to large numbers of people, and if you insist on perfect,
> dublin-core-compliant taxonomies for all data, the rest of the world
> is going to pass you by, even if in some far-off day you might catch
> up with them.
Agreed. And TBL has often made the point that he's not looking for
logic; a Semantic Web has to handle paradoxes gracefully, for example, unlike
most knowledge reasoning systems before it. This is, in a sense, an extension
of that. The advantage of the Web was allowing sloppiness in specifying
location, etc. Some of the putative advantage of the Semantic Web (in
comparison to prior related work) lies in allowing (some) sloppiness in
reasoning. I suggest that Base might offer lessons in the advantages of
allowing sloppiness in semantic design.
More information about the FoRK