RE: Models: Subclassing

["Jim Sermersheim" <jimse@novell.com>]

>From that, I only see that the most subordinate object class must be
structural, and that there can only be one chain (though there can be
multiple inheriitance), there cannot be two completely different chains
(resulting in multiple structural most-subordinate object classes). It
cannot mean that all superclasses are structural as that would nullify
the use of abstract object classes.

Jim

>>> "Pana, Mircea" <mpana@metasolv.com> 1/28/03 10:31:44 AM >>>
X501 section 8.3.2 recommends:
  "An object or alias entry is characterised by precisely
   one structural object class superclass chain which has
   a single structural object class as the most subordinate
   object class."

Doesn't this imply that an auxiliary object class can not
subclass a structural object class?

Thanks,
Mircea.

-----Original Message-----
From: Kathy Dally [mailto:kdally@mitre.org] 
Sent: Tuesday, January 28, 2003 9:58 AM
To: Jim Sermersheim
Cc: ietf-ldapbis@OpenLDAP.org 
Subject: Re: Models: Subclassing


Hi Jim!

There are no restrictions in X.501 on structural and auxiliary classes
being subclasses of one another.  Although it doesn't make a lot of
sense to me, I have seen cases where several structural object classes
have an auxiliary as the superclass.  I can't think of an example of
the
reverse, off hand.

Thanks,
Kathy


Jim Sermersheim wrote:
> 
> Section 2.4 describes object class subclassing, but should express
the
> effect of subclassing (namely that a subclass inherits the MUST and
MAY
> list(s) of superior object classes in its superclass chain, but does
not
> inherit Name Forms or DIT Content Rules)
> 
> Section 2.4.2 states "Structural object classes cannot subclass
> auxiliary object classes." and Section 2.4.3 states "Auxiliary
object
> classes cannot subclass structural object classes.". I can't find
these
> restrictions in X.501--can you point out the reference?
> 
> Jim