Re: BIND and access control lists

The statement that would be consistent with 3744 would be
"BIND is the same as MOVE with respect to the DAV:acl property". 
(I prefer to state it that way, rather than copying the
paragraphs from 3744, since that allows us to modify
DAV:acl behavior in 3744bis without having to revise the
binding specification).

If nobody objects to that behavior, then I am happy to add
that sentence to the BIND specification.

If anyone objects to that behavior, then that is an issue
against 3744, and then I believe the BIND specification
should remain silent on the topic, and the issue should be raised
against 3744 for resolution.

Cheers,
Geoff

Joe wrote on 02/03/2005 02:07:52 AM:

> 
> Trying to predict what the IESG might latch on to during their last 
> call.
> 
> I understand that DAV:acl is a property, which applies to the resource, 
> not the URL.  The suggestion has been made that MOVE has the same 
> issue, but 3744 gives explicit requirements for MOVE and COPY, which 
> differ slightly.  I don't know the history of 3744, but I assume that 
> there needed to be text for section 7.3 and 7.4 since it wasn't obvious 
> what would happen to acl's in the face of those operations.
> 
> I could imagine an IESG member who knew about the entire suite of 
> protocols asking for the draft that added a new operation specifying 
> how that operation fits into that suite.  I would expect that a line 
> saying "BIND is the same as MOVE with respect to acl's", "BIND is the 
> same as COPY with respect to acl's", or a paragraph on the same lines 
> as 7.3 or 7.4 from 3744 would be enough to answer that question.
> 
> -- 
> Joe Hildebrand
> 
>