[
Lists Home |
Date Index |
Thread Index
]
The notion that authority and credentials are themselves
amenable to classification accords well with what I have
read. Typically, classification for security, compartmentalizes
information for viewing by information type and viewee type.
Authorities are classifiable for operations in IETMs using
novice through expert levels, etc. The IETM DTDs used
attributed metainformation for this.
I may have to break down and learn RDF. :-)
len
From: Uche Ogbuji [mailto:uche.ogbuji@fourthought.com]
> True as long as the issues of authoritative
> classification are resolved as well. Then
> the notion of subscribing to a classification
> service becomes more attractive. Credentials
> count. When dealing with a business, one
> qualifies the contenders. That is something
> the DAML+OIL language could be used for in a
> web service.
Yes, and this is one of the things that I think distinguishes our closed-world success story from scaling this concept to the Web. In many cases we were able to solve "web of trust" type problems through policy.
However, I think that once you simplify the problem to classifications, you also simplify the authority problem, because the level of authoritativeness just becomes a meta-classification, and can be operated on using the same class of automata as those that operate on the rest of the ontology. I think that if such ontologies can scale to the Web at all, that they will be able to handle classifications of authority.
|
