RE: Re[2]: [higgins-dev] Display and schema information in IdAS API

["Paul Trevithick" <paul@xxxxxxxxxxxxxxxxx>]

Valery wrote
> 
> Jim,
> 
> I agree that it would be a good thing to have this kind of information
> on the idas.model API level. The question is where and how could we
> define it.
> 
> With regards to _required_ and _optional_ attributes it looks like
> more or less simple case - we could just use owl:minCardinality and
> owl:maxCardinality restrictions in the context schema to define
> whether some type of attribute is required or not.
> 
> Unfortunately, it is not so clear for me how to be with the matching
> rules.
> 
> My first intention with display data on attributes was to use some
> kind of metadata but when I knew more about OWL DL restrictions I
> realized that it is impossible. OWL DL doesn't allow to add extra
> information to the property axiom but it is exactly what I was need to
> add metadata to attribute's definition.

As a word of background, there is already a growing mass of evidence that we
may want to do away with our current, self-imposed restriction of using
OWL-DL and allow ourselves to use the more permissive OWL-Full instead.
Relaxing to OWL-Full would allow us to (i) cleanly express the semantics of
LDAP classes (e.g. AUXILIARY) that we've not been able to do, (ii) use XML
Schema types to define the data values of Attributes. The downside to using
the more permissive and expressive OWL-Full is that we will not be able to
use off-the-shelf reasoners (performing inference, etc.) to answer queries
about schemas. My approach has been to see how far we can go with OWL-DL
before cutting over. Because there is no return once this decision is made.
I've been thinking for some time that we'll want to move to OWL-Full, but
holding to OWL-DL in the meantime helps bring the limitations of OWL-DL into
clear focus.