Teleconference minutes -- November 1, 1999
Jim Whitehead (ejw@ics.uci.edu)
Mon, 1 Nov 1999 12:04:09 -0800
From: Jim Whitehead <ejw@ics.uci.edu>
To: ietf-dav-versioning@w3.org
Date: Mon, 1 Nov 1999 12:04:09 -0800
Message-ID: <NDBBIKLAGLCOPGKGADOJGENJCHAA.ejw@ics.uci.edu>
Subject: Teleconference minutes -- November 1, 1999
Delta-V Teleconference
November 1, 1999
Present: Jim Amsden, Chris Kaler, Geoff Clemm, Jim Whitehead, Brad Sergeant,
Jeff McAffer
Some brief discussion on the access control protocol. Jim Amsden wonders why
more people haven't been interested in this protocol -- lack of an
interoperability protocol seems to be a significant missing link in WebDAV.
Chris noted, from his DCE experience, that this is a very hard problem.
Geoff noted that you can perform a lot of useful work without them.
There was some discussion on the process for issuing drafts. There were
some concerns that some items were flip-flopping from revision to revision,
and hence would be difficult to read by external observers. Agreement that
there needs to be working group discussion when adding/removing contentious
items from specificaitons as they're being written.
Discussion of which items should be discussed in the Delta-V WG meeting:
- property resources and property collections
- MRESOURCE
- label vs. URL segments
Changed our minds from last week, and decided to *not* give a presentation
on goals, since that's been done several times already. Discussed whether we
should have a presentation on detailed semantics of the protocol. Suggestion
that we do a scenario walkthrough. Use the scenario to drive discussion of
issues -- for example, when creating a resource comes up in the scenario,
use this to discuss the MKRESOURCE issue. Geoff Clemm volunteered to develop
a scenario for discussion during the WG meeting. This scenario will cover
the core semantics of the protocol (check-in, check-out, labels,
workspaces).
Chris Kaler issues from reading through the draft.
* 2.1 we say you put something under version control by checking it out.
There is an open issue on how exactly we put items under version control.
However, if we look at the protocol, CHECKIN and CHECKOUT have preconditions
that require versioned resources -- this is a contradiction.
* We need to use the ISO datetime specification throughout the spec., not
the HTTP date format.
* We discuss impact of versioning on existing methods, but we don't mention
the impact of versioning headers on versioning methods (for example,
Ref-Target).
* Should lock have a side effect of doing a checkout?
* Checkout and Checkin take the exact XML element that is used for
PROPPATCH. We should use a separate one for checkout and checkin, to
separate the namespaces. Do we also want a checkpoint operation (checkin,
but keep check-ed out).
* Echoing. If someone specifies a target selector, the response should echo
back the target selector used.
*** End of teleconference ***