W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2003

Bind issues

From: Lisa Dusseault <lisa@xythos.com>
Date: Tue, 18 Nov 2003 11:41:27 -0800
To: "'Geoffrey M Clemm'" <geoffrey.clemm@us.ibm.com>, "'Julian Reschke'" <julian.reschke@gmx.de>
Cc: "'Wallmer, Martin'" <Martin.Wallmer@softwareag.com>, "'Webdav WG'" <w3c-dist-auth@w3c.org>
Message-ID: <01c001c3ae0b$f4ebe740$75c990c6@lisalap>
On 8/4/2003, I asked about PROPFIND results in presence of bindings.  I
don't know that
that has been clarified.
On 3/8/2003, I proposed language for clearer requirements on DELETE
behavior. This applies to MOVE too.  The discussion started earlier,
3/4/2003 or before.
I believe I asked for more lock requirements -- what the server MUST do when
a client locks a binding.  I haven't found that email yet.
It's quite possible that the bind authors believe they've already dealt with
these issues.  If so, I'm unaware of the resolutions.

-----Original Message-----
From: Geoffrey M Clemm [mailto:geoffrey.clemm@us.ibm.com] 
Sent: Tuesday, November 18, 2003 5:29 AM
To: Julian Reschke
Cc: Lisa Dusseault; 'Wallmer, Martin'; www-webdav-dasl@w3.org
Subject: Re: "URI properties", Was: SEARCH for displayname

Like Julian, I have no memory of Lisa raising this issue on 
the mailing list (at least, not since the activity on the BIND 
protocol was resumed a year ago). A pointer to that email would be
since both Julian and I work very hard to maintain accurate issue 
lists and to make sure that all issues that are raised on the mailing 
list are addressed.  There was a discussion in the design team 4 or 
5 years ago, about whether the spec should leave the question of where 
the state of a binding resides up to the server (i.e. is it in the 
parent collection, the child resource, or both), and the consensus 
was that effective interoperability required us to specify where 
this resides, so that access control and locking behavior could be 
defined (i.e. was the removal/addition of a binding controlled by 
a lock/acl on the parent collection, on the child resource, or both). 
The choice of the parent collection was motivated by the use case 
of two bindings to the same resource in a single collection, and by 
the use case of different collections needing to give different names 
to a single resource. 


Julian Reschke <julian.reschke@gmx.de> wrote on 11/12/2003 11:00:00 AM:

> Lisa Dusseault wrote:
> > I do plan to raise these issues more generally for bind.  However, I 
> > have also raised them in the past and they do not show up on your list.
> Well, the please re-raise them with a pointer to the original mail on 
> the mailing list.
> > Sorry if my replies are brief to the point of bluntness; I'm trying 
> > to do a bunch of IETF coordination and cross-group work this week and
> > attend other WG meetings at the same time.  I will try to understand
> > the context in which you're saying how things are and must behave.
> Good.
> > In return, please respect that I am challenging the model assumptions
> > I see developing here and in bind discussions.  I hope it's not too
> > late for us to have open minds about how things are defined.
> The model for BIND has been developed in 1999 and has never changed since.
> Julian
> -- 
> <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
Received on Tuesday, 18 November 2003 14:43:50 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:28 UTC