RE: Advanced collections and ordering

Thanks for jumping in, Jim.  I did actually make changes to the spec based
on Max's comments.  I hope that version 3.2 of the collection spec is
clearer.  It's on the Web at
http://www.ics.uci.edu/~ejw/authoring/collection/dt/CollSpec032.txt.

I should say that even though 3.2 include some support for server-maintained
orderings, there is no consensus even on the design team that it should stay
there.  So it may yet be removed.  Whether it stays or goes, making the
distinction between client-maintained and server-maintained is important, so
at least an explanation of the difference and a statement about what we
support will stay.

--Judy

Judith A. Slein
Xerox Corporation
jslein@crt.xerox.com
(716)422-5169
800 Phillips Road 105/50C
Webster, NY 14580


> -----Original Message-----
> From: Jim Davis [mailto:jdavis@coursenet.com]
> Sent: Wednesday, April 21, 1999 3:21 PM
> To: w3c-dist-auth@w3.org
> Subject: RE: Advanced collections and ordering
> 
> 
> Not sure anyone ever answered you Max
> 
> At 06:56 PM 3/9/99 +0000, Max Rible wrote:
> >At 14:33 3/9/99 +0000, Slein, Judith A wrote:
> 
> >>The server is required only to insure that, for collections that are
> >>ordered, every collection member appears in the ordering 
> exactly once and no
> >>resource that is not in the collection appears in its 
> ordering.  Beyond
> >>that, it's up to the client to insure that the ordering follows the
> >>semantics identified by DAV:orderingtype.
> >
> >Are you saying that PROPFIND will always return elements in arbitrary
> >order and that the client should take the rules and do the 
> sorting itself?
> >This seems at odds with the collection protocol.
> 
> No, the server preserves the order that the client established.
> 
> In the base WebDAV spec, the state of a collection includes the set of
> members, but not the order of elements in the set.  Hence a 
> server could
> return a different (random) order each time.  In an ordered 
> collection, the
> state includes not only the set but also the order.
> 
> The server does not know the meaning of the order, and does nothing to
> enforce it.  It just promises not to randomize it.
> 
> Is this clear now?
> 
> If you think Judy's language was confusing please suggest 
> better language
> that would have made this clearer.
> 
> Best regards
> 
> Jim
> 
> please reply to jrd3@alum.mit.edu, despite the Reply-To 
> address in the header.
>

Received on Wednesday, 21 April 1999 16:31:34 UTC