Re: WG Last Call on URP I-D

At 08:11 AM 2001-09-05, Richard V Huber wrote:
>Yes, that is why I said that I originally thought a reference to the
>type would include all subtypes for a delete. But 2251 does not go
>into detail on what it MEANS to be "treated as a "subtype of the
>attribute type without any options".

I believe it means that 'cn;foo' is to 'cn' as 'name' is to 'cn'.
Hence, deleting 'cn' would NOT delete 'cn;foo'.  I believe 2596
is consistent with this interpretation.

>Is LDAPBIS going to reissue 2596?

It's not a charter item.  I would suspect the authors would
revise this based upon clarifications to LDAPv3 produced
by LDAPbis and language tag clarifications already published.

I note that LDAPbis is considering the impact of changes it
proposes upon this and other extensions.

>How do we deal with the inconsistency?

Well, the primary inconsistency is that we currently have
two different semantics for options:
        1) most options act as subtypes
        2) transfer options (;binary) don't act as subtypes

I've heard two different suggestions to deal with this:
        a) detail two kinds of options
        b) axe transfer options

Currently, LDAPbis is attempting a).

If you have comments in this area, I suggest they be directed
to the LDAPbis list.