[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: models: objectClass;option



At 07:44 AM 4/16/2004, Hallvard B Furuseth wrote:
>Should [Models] say anything about objectClass attributes with
>options?

Well, should [Models] say anything about subtypes of this or
various other attributes.  I think not namely because one
could devise an extension where that made sense.  For
sense, one could have 'objectClass;auxiliary'
'objectClass;abstract' and 'objectClass;structural' which
separately listed classes of each kind, if requested
that way.  I don't see much problem with that.

I do see a problem, in general, with extensions which
are not truly optional.  While I think that's mostly a
topic for an extension guidelines document, I think it
would be wise for to make a statement in this area in
an appropriate place.
        LDAP is an extensible protocol.  Extensions are
        expected to be truly optional.

>I can't say it worries me, but maybe someone else cares.
>
>That is, MAY/SHOULD/MUST [NOT] the implementation treat an object class
>attribute with a recognized option as a normal object class unless the
>option description says differently?
>
>Example:
>  dn: l=foo,o=mysil
>  l: foo
>  objectClass:         locality
>  objectClass;lang-no: person
>  objectClass;lang-en: country
>
>-- 
>Hallvard