[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[higgins-dev] Higgins Dev Call September 7 at 12pm ET

- Heather, Raj, Mike, Tony (IBM)
- Jerry (Sun)
- Pat, Tom, Jim and Dale (Novell)
- Brian (Serena)
- Paul, Mary (SocialPhysics)
- Uppili (Oracle)

Component architecture 
- http://wiki.eclipse.org/index.php/Core_Components
- Discussion of latest change in v22: splitting I-Card Registry out as a
separate component from ISS. 

- Group is feeling urgency of getting the IdAS API to a first cut state
- We may need to have two versions, one more stable, one latest and greatest
- Need higgins.owl done (Paul promised by Friday 9/8)
- The uncertainty about attribute handling is holding up finishing that part
- Agreement: work with Greg on Novell proposal for first cut of IdAS API
- We'll probably create our own interfaces influenced by Jena

Architecture change related to attribute->claim mapping
- There was a discussion of another architectural point...
- Most of us agreed that IdAS should only know about attribute namespaces,
not claims
- Paul thought that the Token Issuer (or its Token Provider) was the place
to do attribute->claim mapping
- Mike explained why it should intead be in the I-Card Provider
- Raj agreed, but said it should also be a separate logical component
- In the end...
  o We mostly agreed that we may need to support both a "push" and a "pull"
  o "Push" meant that the I-Card Provider could be provided a reference to
the claim values needed as part of the RST. (It would in turn push these to
a Token Provider)
  o "Pull" meant that a Token Provider would consume the claim values from
the I-Card provider (well, from the attribute->claim mapping used by the
I-Card Provider)
  o We agreed that attribute->claims mapping was worthy of being a separate
  o Paul will prepare a v23 diagram for further review/discussion

OWL editors
- discussion of the lack of a perfect open source OWL tool
- Paul uses SemanticWorks2006
  - not open source 
  + adequate doc
  + good functionality (including diagramming)
  - poor OWL RDF writer (produces very verbose)
- Paul has tried (open source) Protoge with OWL plugin but it crashed and
lost work
- Paul has tried the SWOOP tool from MINDLAB (UofMaryland)
  + open source
  + OWL-focused, browser-based UI
  - arcane UI, none/poor doc

LDAP to OWL issues
- We discussed the fact that in OWL an instance is considered to be a member
of an OWL class so long as it satisfies the constraints on required,
optional properties, values, etc. but (unlike, say, Java) if the instance
also includes other properties, it is still considered an instance of the
class. The reason is that in OWL Classes are named sets of instances, not
"definitions" of the instances. A subtle difference.
- Tom explained that the literature related to LDAP to OWL/RDF was shallow
- Paul suggested he talk to Mark Wahl (mark.wahl@xxxxxxxxxxxxxxxxxxxx) who
is creating an LDAP to OWL extract/generate tool
Higgins F2F at DIDW
- Mary: Higgins meeting will be Tuesday afternoon, still waiting for a room
- We're hoping to see if Kim could make it to this meeting