SWBP Community Feedback

Hi all,

I recently attended the US Army KM Conference 
(http://www.afcea.org/doim2004/) and one of the Top Quadrant Semantic 
Web Seminars (http://www.topquadrant.com/seminars/topmind.htm).  This 
message summarizes feedback on the direction of the SWBP that I 
received there from attendees.


Overview of SWBP Direction
--------------------------

SWBP documents expected by SemWeb-aware practitioners are general 
guidance on practical issues, as well as implementation guidance.  The 
consistent feedback that I received was that the documents published to 
date have been "too geeky".  I think that both types of documents have 
their place, but we should strive to produce both types.


Desired Documents from SWBP
---------------------------

1)  Provide a document mapping XML Schema to OWL.  This was requested 
by several people, notably Mike Daconta (http://www.daconta.net - now 
the Metadata Program Manager at the US Department of Homeland 
Security).

2)  Provide guidance on how to migrate generic taxonomies (including 
UML) to OWL.  Specifically, how can one develop a private ontology 
which makes up for OWL's lack of 'PartOf'?  I am told by Jim and Guus 
that this is straightforward, but many people are under the impression 
that since OWL does not support 'PartOf' natively that it is not 
possible.  I think we should correct that impression.

3)  Provide guidance on how to represent common ontologies, such as 
units & measures and latitude/longitude.  This is a real issue for a 
lot of people.  Over 15 people in one room were independently 
developing ontologies that represented lat/log.  They seem to want 
*guidance*, not a canned solution.

4)  Provide guidance on how to represent data in RDF/XML instead of 
some other XML vocabulary.  Fannie Mae wants to publish data in XML and 
Eric Miller suggested that they use RDF/XML.  That one company could 
drive an entire industry to doing so, but they need guidance from 
someone to help them.  They want some guidance on how to represent 
their data this way, since triples scare them a bit.


Usability
---------

Several people spoke to me about how to actually *implement* the 
Semantic Web use cases.  Perhaps we could look at the existing SemWeb 
use cases, expand them into full scenarios and discuss how they could 
be implemented with existing (or near-existing) technologies.  This may 
be best spread over the existing TFs?


Regards,
Dave
--
David Wood
CTO, Tucana Technologies, Inc.
http://www.tucanatech.com

Received on Monday, 13 September 2004 16:10:10 UTC