[Person-ontology] Introduction and original goals

[Paul Trevithick]

My name is Paul Trevithick. I started what has become the Higgins project in
2003 because I wanted to create a dashboard that provided a unified, virtual
view of all of the "me"s (including my relationships with others) that were
(and still are) contained in many separate web and enterprise "silos". Since
every silo had its own schema, an abstraction layer was needed. Since I knew
that context was critical, I wanted to be able to preserve the
representation of my separate identities in different contexts and yet be
able to link them across contexts. I looked around and didn't see any
off-the-shelf technology that leveraged a rich enough data model. I new that
to virtualize and unify I'd need to move from the syntactic to the semantic
realm.  I didn't want to have to invent a new data model and keep updating
it, so I selected RDF as the foundation.  The Higgins project has grown from
one developer to many. We have established an abstract ontology
(higgins.owl, affectionately know as HOWL) that all "context providers"
(adapters to legacy databases of social networks, profiles and other kinds
of identity data) must be based on.  Towards the goal of interoperability,
this is an initial baby step, but far from sufficient. Each context defines
its own higgins.owl-based ontology and so mapping is still required across
them. The Higgins project documents its data model here:
http://wiki.eclipse.org/index.php/Higgins_Data_Model including motivation,
design goals, etc. 

 

Enough about Higgins. 

 

The next baby step in our industry would be the creation of a website where
people developing these different ontologies could advertise theirs and see
others. This might help prevent the current state of affairs where folks
don't have an easy place to even see what others are doing, so everyone just
creates their own ontologies willy nilly. Mark Wahl and others had similar
thoughts so http://idschemas.idcommons.net <http://idschemas.idcommons.net/>
was born (aka http://identityschemas.org <http://identityschemas.org/> ). In
addition to Mark's Schemat efforts, there is now the beginnings of a
"community dictionary service" (http://cds.idschemas.idcommons.net
<http://cds.idschemas.idcommons.net/> ). It is very much a work in progress,
but could also contribute to interoperability.

 

And now, some truly intrepid folks (i.e. those on this list) are willing to
try to work on a common, universal ontology about people identity, profiles,
relationships, etc-everything about a person. Everyone says this is
impossibly hard. Every previous attempt has failed. Moreover even if we can
do it, it will never be adopted. Perhaps. But even incremental progress
would be progress indeed. 

 

-Paul

 

PS: This list is about more that just Higgins. But here's a final
Higgins-related note. Higgins.owl introduces a formalism that allows what we
call "metadata" statements to be made about "attribute" statements. This is
very awkward to do in RDF because reification is so ugly, but in the world
of identity management it is a critical requirement. If folks on this list
seek to improve/replace the foundational higgins.owl layer, the replacement
must also support this ability to make statements about statements. Also,
every step of deviation from the semantics of higgins.owl requires a change
to the IdAS SPI and API-a heavy cost that must be weighed.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://idcommons.net/pipermail/person-ontology/attachments/20071007/4765d357/attachment.htm