Re: [higgins-dev] Re: Proposed Version 1: Higgins based LDAP OWL ontolog

["Tom Doman" <TDoman@xxxxxxxxxx>]

I should have been more clear on my purpose.  The main purpose of the proposal was not to indicate what schema we thought was important to include.  Rather, it was to propose the generation scheme used to generate an OWL ontology from LDAP schema.  In fact, it wasn't my purpose at all to propose what schema should be present.  Thus I wasn't very worried about what our test server was providing us.  Our Higgins LDAP Context Provider will not be able to count on what schema any given LDAP server will provide, it will simply dynamically generate an OWL ontology based on what's it's given.  I made the proposal based on the assumption that others might not represent LDAP schema in OWL the way we have chosen to do it.

A couple of points FWIW:
1. I can point my CP at some other LDAP servers and planned to do that anyway because I want to verify it works correctly on servers that'll provide a lot more schema.  I'll send out the results so whoever is interested can have a look.  Could you point me at a public LDAP server that has the schema you'd like to see in it?  I'll generate an ontology based on it and send it out.
2. My syntax mapping table currently only covers RFC4517 syntaxes.  I plan to expand it to include other well known syntaxes for more explicit readability.
3. My syntax mapping table is currently hard coded but I plan to make it part of our CP configuration file so it can be expanded.

Tom Doman
Novell Inc.

>>> Mark Wahl <Mark.Wahl@xxxxxxxxxxxxxxxxxxxx> 10/2/2006 10:23 AM >>>
Tom Doman wrote:
> Yep.  I'm exercising "garbage in, garbage out" on that one for now. 
 > The LDAP server I'm getting that from is a simple test server that
 > doesn't even validate entries against the schema anyway.

The concern I have is that in your post you wrote that

 > At least, we believe this is complete enough to call "version 1".
and that "I'll propose this to the Identity Schemas group".

I would have expected that a version 1 of an ontology proposed as the
Higgins base LDAP OWL ontology should use as a basis for definition
either the most recent standards-track LDAP schema, as defined in the
IETF LDAP RFCs, or the schema elements published in the IANA LDAP
registry, and not just the LDAP schema of a specific implementation
which differs from both of the above, since there is a consensus-building
process for whether a schema element is present or not in your server
or not.  It is difficult to review a ontology definition when it is not
clear whether certain features are deliberate design choices, or are
due to limitations of the LDAP schemas used as the source for your
ontology.

For example, I note that when comparing your file against RFC 4512
and RFC 4519, the schema is also missing the extensibleObject,
dcObject, uidObject classes, the supportedFeatures attribute, etc.
It would also be useful to consider having the schema defined in
RFC 4524, such as account and document, included.

Mark Wahl
Informed Control Inc.
_______________________________________________
higgins-dev mailing list
higgins-dev@xxxxxxxxxxx 
https://dev.eclipse.org/mailman/listinfo/higgins-dev