Re: ;binary and (supertypes of an attribute description)

["Timothy Hahn" <hahnt@us.ibm.com>]

Jim,

Such a categorization sounds like an interesting idea.

However, we should phrase this is such a way that all future attribute descriptions that are created/defined MUST declare whether they are "subtyping options" or NOT.  (I don't think it's required that we necessarily know that they are "transfer specifier options" - it seems sufficient that we would know that they are NOT to be considered for "subtyping").

Regards,
Tim Hahn

Internet: hahnt@us.ibm.com
Internal: Timothy Hahn/Endicott/IBM@IBMUS or IBMUSM00(HAHNT)
phone: 607.752.6388     tie-line: 8/852.6388
fax: 607.752.3681

Sent by:        owner-ietf-ldapbis@OpenLDAP.org

To:        <ietf-ldapbis@OpenLDAP.org>, <Kurt@OpenLDAP.org>
cc:        
Subject:        Re: ;binary and (supertypes of an attribute description)



I'd like to find a general way in the protocol document to make a distinction between "transfer specifier options" and "subtyping options". Once that is done, I'd like to make the "An AttributeDescription with one or more options is treated as a subtype" statement apply ONLY to "subtyping options".

Are there reasons not to do this?  If so, I'd have to vote that each "transfer specifier option" specify whether or not it is an exception to the subtype rule (more explicitly than is currently being done).

Jim

>>> "Kurt D. Zeilenga" <Kurt@OpenLDAP.org> 3/14/01 9:56:14 AM >>>
The essence of this issue is application of:
  An AttributeDescription with one or more options
  is treated as a subtype of the attribute type
  without any options.

to an AttributeDescription such as userCertificate;binary
and the description that the binary transfer option.

That is,
  userCertificate;binary is treated as a subtype of
  userCertificate.

and section 4.1.5.1 implies that userCertificate;binary
is NOT to be treated as a subtype of userCertificate.

Ugh.

Kurt