Thursday, April 12, 2007


This idea came up in the context of a discussion regarding our summer boot camp for incoming students. We are just trying to think of something for them to do that will teach them about the semantic web but that might also lead to something interesting. Right now, all roads seems to be leading me to identity.. Before reading on, you may find the links here useful - particularly Dick Hardt's video presentation.

In the context of data integration, establishing identity is a crucial first step (see LSID papers here for more explanation In the wild world of WebX.0, the most critical entities in need of unambiguous identification are people. If we had a consistent, coherent system for establishing who (in a global fashion) is the Mark Wilkinson attempting to buy that house - or attempting to edit my wiki - or answer my icapturer question - or is the author of which paper, then we would have at least two obvious, immediate advantages; 1) when, as open-input application developers we need to make decisions about who to trust, we can use the meta-data associated with the putative person on the other end of the wire to do so, 2) we would have a coherent, once again global, way to give credit where it is do. The real-world would not work the way it does if it were anonymous.. an LSID-based identity management system for the Web might just make things on the Web a bit more real.

Specifically we would need to:
a) build the relevant LSID bits (the resolver, lsid assigner , etc.)
b) Provide a website where people could sign up for their own LSID
c) Use the LSIDs as the identity nodes for the Wilkinson Lab Semantic Web
d) Define an extension to FOAF (if required) that lets us make statements about trust.. and reason over them
e) Use FOAF and our own trust ontology to define a consistent structure for the RDF meta-data associated with each LSID. (Add all the usual who knows who, who wrote what bits as well..)
f) Make a demo website - maybe a wiki for the lab since we need one anyway - that utilizes the the personal LSIDs for semantic user-validation
g) read the book "True Names" by Vernor Vinge which described the importance of this concept for cyberspace 26 years ago.