Editor's note:  These minutes have not been edited.

		Access and Searching of Internet Directories WG Meeting
				Meeting Minutes
			Wednesday, December 11, 1530-1730
				Reported by: Tim Howes

- Agenda review/changes

	The proposed agenda was slashed quite a bit, with some items
	punted to the list, in an effort to make room for LDAPv3,
	which was anticipated to require a lengthy discussion. Items
	dropped were: pgp draft (to the list), domains draft (discussed
	already in IDS), cip and ldap discussion (discussed already
	in FIND). Items cut down in time were: whois++, rwhois.

- application/directory MIME type drafts

	- application/directory framework

		Tim reported that a new application/directory framework
		draft had been produced which addressed all outstanding
		comments received. A brief discussion revealed several
		more issues with the draft that people raised.

		These issues were:

		- Example is wrong in how it does line breaks.

		ACTION: Tim to fix this in the draft.

		- Using MIME vs. BEGIN: END: sentinals to carry
		  multiple parts.

		ACTION: Discussion to take place on the list.

		- Change the "proto" parameter to "context"

		ACTION: Tim to change this in the draft.

		- Reference to RFC 1123 time/date formats should be
		 change to reference an I-D describing the ISO 8061
		 time/date format. Chris Newman volunteered to write
		 up this draft.

		ACTION: Tim to fix references in the draft.

		ACTION: Chris to write up the 8061 draft.

		- Ned Freed and Kevin Jordan both had comments that
		  they agreed to send to the authors and/or bring
		  up on the list.

		ACTION: Ned and Kevin and others with issues to bring
			them up on the list, and/or to give feedback
			directly to the authors.

	- vcard profile

		Frank Dawson reported that the vcard profile draft had
		been revised to address all known problems and issues
		raised at the last meeting. One additional issue was
		raised at this meeting: the use of MIME media types
		for audio and photo types. The group felt this would
		be better than devising a new scheme.

		ACTION: Frank to revise the draft to reference the
			MIME media type registry.

- WHOIS++ drafts

	New WHOIS++ drafts have been produced which address
	various problems found during implementation of the
	drafts. These include:

	- Multi-language handling

	- Separate INCHARSET and OUTCHARSET parameters.

	- New templates for X.509, PGP, etc.

	ACTION: Tim to ask the ADs to re-issue these
		documents as proposed standard.

- RWhois

	Network Solutions is working on a meta-directory service
	that will map organization and domain names to directory
	services. Version one supports RWhois. The next version
	will support more general access from other protocols and
	the ability to refer to arbitrary directory services via
	URLs.

- LDAP API

	Tim and Mark produced a new draft updating RFC 1823,
	describing the LDAP API. The updates include preliminary
	support for the changes expected in LDAPv3, support for
	threading, better data encapsulation, etc.

	The group discussed the future of this draft, whether
	they wanted it brought within the working group, and
	if so, what track should it be put on (standard, informational,
	experimental). The group consensus was to bring the
	draft into ASID so it would get the careful review it
	deserves. The group decided to try to push the draft
	along the standards track initially, with informational
	as a fall-back.

	An issue was raised about draft ownership and perceived
	credit, should the draft become an informational RFC.
	The concern was that an informational document that was
	essentially the product of a single company rather than
	the working group, not be presented as the work of the ASID
	group. Only if the group has consensus on the draft and
	feels it has had sufficient input to it, should the draft
	be advertised as a product of the ASID working group.

	ACTION: Tim to re-issue the next version of the draft
		to the working group.

- LDAPv3

	The LDAPv3 discussion began with Mark Wahl summarizing
	the outstanding issues with the current drafts. These
	issues and others raised during the first part of the
	meeting were:

	- The relationship between SSL authentication and the
	  LDAP Bind operation needs cleaning up.

	- Compliance - What does it mean to be LDAPv3 compliant?
	  The current drafts are not clear.

	- Normalized matching - Do we really want to make this
	  optional, as stated in the current draft?

	- Paged searching - When can the server discard result
	  sets from searches? Some discussion that this no longer
	  matters, since each paged search request now contains
	  enough information to reconstruct the original search.

	- Bind as DN w/out password - The semantics of this
	  operation need clarifying.

	- Mapping onto LDAPv2 - Needs clarifying.

	- Mapping onto DAP - Needs clarifying. Should this be
	  throughout the document, in a separate document, or
	  in an appendix?

	- X.500 93 subentries on search - This is believed to
	  be covered by doing an explicit search for the proper
	  object class.

	- Relationship of the X.500 93 contexts feature and
	  the current multi-language support - This needs to
	  be reexamined and clarified to see if 1) there is
	  more valuable stuff we can steal from X.500 and 2)
	  there are small changes we can make to be more
	  compatible with X.500 93 without increasing complexity.

	- Additional SASL mechanisms - Should we define some.

	- X.500 97 user requirements - [[can someone explain
	  the issue here?]]

	- Mapping of strong authentication - How does this
	  map onto DAP? What does it mean?

	- General direction of LDAPv3 - Some people feel it
	  is too complex.

	- LDAPv2 revisions - should this be progressed or dropped
	  in favor of LDAPv3 entirely?

	- LDAPv2 coexistence strategy - We need one.

	Discussion very shortly centered around two related topics:
	The future of the LDAPv2 drafts, and the general feeling
	that the current LDAPv3 proposal represents an overly complex
	revolutionary rather than evolutionary change to LDAPv2.

	Harald emphatically stated that LDAPv2 could not be progressed
	past draft standard since it has the following known
	fatal deficiencies:

		- No referrals

		- No internationalization support

		- Broken handling of certificates

		- Generally insecure password-based authentication

		- No extensibility mechanism

	There was much discussion about the best approach to take
	to fix these deficiencies in LDAPv3. The debate soon
	centered around two options, the final form of which are
	presented below:

	1) Start with the LDAPv2 RFCs and add support for referrals,
	   i18n, extensibility, and better authentication. Fix
	   the broken certificate support.

	2) Start with the LDAPv3 drafts and do a brutal feature
	   review and cut with the following criteria: Anything
	   that's in must solve one of the problems above. Other
	   features to be added later via the extensibility
	   mechanism.

	A third option that involved bludgeong Harald into letting
	the group progress LDAPv2 as is was quickly dismissed,
	much to Harald's relief.

	There was much debate and an initial straw poll showing the
	room pretty evenly divided between the two options. After
	much "concensizing", the group actually came to a miraculous
	concensus view that approach 2) was the way to go, provided
	there was a way to ensure that the feature review and cut
	would actually happen.

	Tim proposed and the group agreed that a small group of
	motivated volunteers should be tasked with going off and
	doing the feature review and cut, which would then be
	brought back to the group. The group agreed that this
	task must be completed by January 31, 1997.

	ACTION: Tim to organize the feature review and cut posse.

- Any Other Business

	The meeting concluded with consensification, almost on time.
	The next ASID meeting will be in April in Memphis, TN, USA.