Menu

Projects:

  Higgins

  Identity Gang  

  Identity Schemas

SocialPhysics.org

SP Blog

Paul's Blog 

Introductions by Attendees
--------------------------
Mary Ruddy - Higgins, interests: interop
Tom Doman - Novell, Higgins, Bandit; interests: interop
Iain Henderson - customer information consultanty, privacy, personal inf. mgmt
Johannes Ernst - Netmesh - common schemas since 1993, interop
Jim Sermersheim - Novell, Higgins, x500 directories
Joaquin Miller - Netmesh, got interested at Cargill
Dick Hardt - Sxip Identity - protocols for moving around identity data
Barry Ferg - Sxip - working on protocol schemas
Drummond Reed - Cordance, XDI.org, OASIS - XRI, XDI data interchange, data interchange schema, dictionary representation is in XDI TC
Bob Morgan - UofWashington, Internet2 - working on interop for several years in identity systems, eduPerson schema, working with Europeans, worked with Library of Congress communities, main problems with attributed-based sign-in is schema mapping, worked on USperson for Fed Government
Paul Trevithick - Higgins, interested in interop
Dale Olds - Bandit and Higgins project

The following are paraphrases of what folks said (and much is missing):

- Paul's suggestion for an agenda: walk thru and review http://identityschemas.org

- Johannes: When I first saw the CardSpace schema and saw "surname" for the umpteenth time. This happens partially because there is no place to go to see what's been don. If we only accomplish creating central place, we will have done something. We want to set our goals really low and get something done.

- Bob: As I mentioned at the last open space. The barriers to re-using schemas and not strictly technical. Our experience with our European partners shows examples of this. Mapping is always necessary.

- Johannes: First we need the ability to find what's there already. Point, don't reinvent.

- Tom: starting with the just the ability to express in a std way what's already there. E.g. OWL.

Walking through the wiki sections...

Purpose
-------
- No discusssion

Problems we're trying to solve
------------------------------
- Some discussion that a reputation system in the long run may be needed. Probably not an immediate problem to tackle.

Deliverables
------------
- Bob: we should add documentation as a deliverable: "how to think about using schemas, "how to apply what's on this site", etc. Guides, etc.

IP Policy
---------
- Bob: policy about the service we're trying to build but not the process.

- Drummond: this is at odds with what is on the id2 wiki

- Drummond: we need add an IPR statement regarding the process (as opposed to the product)

- Drummond: is the consensus that this is an open, transparent and that noone is going to assert IP rights over what happens in the process.

- Bob: it is common for groups to say that anything sent to this list is open for use, etc.

Implementation
--------------
#1 Schema description language
- no discussion

#2 Base Ontology
- Mark has suggested that the use of a base ontology (e.g. higgins.owl) be optional instead of required. I think there was rough agreement on Mark's suggestion.

- Bob: what are the requirements for contributions? We should list these requirements clearly and explicitly

#3 Mid-Level Ontologies

- Marks point is similar to #2

- Johannes: what's folks reactions to OWL?

- Paul: unfamiliar, strange, but after a while acceptance

- Johannes: the Semantic Web is lots of export, little import. OWL is about as good/bad as other options, what can we say about import

- Tom: we haven't much experience with import

- Jim: our experience is so little, we don't know all the variations.

- Johannes: we should look carefully at these import issues

#4 Cross references

- Drummond: if there is a specific benefit from XDI, it would be the XDI dictionaries. What links to this? etc. This could be a synergy here.

- Johannes: mapping between schemas is very difficult, [gave examples of addresses, etc.] Its conceivable to have "full name" in one schema and "first" and "last" in another. How do you map between them.

- Johannes: this is the most challenging area. There are real problems here. Lets try the easy ones first.

Wrap-up
-------

- Johannes: What are the mechanics going forward? How do we make contributions. E.g. we've got an analysis that we've already done of some schemas, where can we contribute it, etc.

- Paul: we'll continue for another week using the email cc list and then the IC commons IdentitySChemas list should be ready. We'll also cut over to another wiki.

- Paul: Though the plan is for a db-backed site, we should see how far we can go on the wiki too.