Re: Moving forward on BIND

Jim Whitehead wrote:
> Joe writes:
> 
>>My guess for BIND is that either Ted or Scott would want 
>>clarification on a couple of the interoperability questions 
>>that Lisa has raised, regardless of whether the answer can be 
>>inferred by a fully-informed reader. 
> 
> 
> Julian Reschke and I just chatted on the phone. My perception is that we
> both agree that it makes sense to discuss the semantics of all DAV
> properties under BIND, UNBIND, and REBIND. There seem to be differences in
> desired behavior between properties whose semantics originated in WeBDAV
> (which are typically defined on the resource), and ones inherited from HTTP
> (like getlastmodified) which depend on representations of the resource. It
> is likely this discussion will lead to modifications to the bind
> specification.
> 
> Julian said he'd take the first whack at summarizing the technical points.

OK,

I have started a summary of what I think is the relation between HTTP 
header data and WebDAV namespace operations (including but not limited 
to those in BIND). Current draft at: 
<http://greenbytes.de/tech/webdav/draft-reschke-webdav-namespace-vs-properties-latest.html>. 
Writing this down makes a lot of sense; let's try to discuss & 
understand the problem first before we argue whether and what BIND needs 
to say about it.

(for those too lazy to read it... :-):

- we can't say anything about properties where we inherit semantics from 
RFC2616 (HTTP), just like Roy F. pointed out a few months ago. This 
includes DAV:getlastmodified and DAV:getetag.

- we could try to say a few words about those properties that are 
defined by RFC2518 and friends (although I'm not convinced it's worth 
the effort).

There are a few issues in BugZilla that have been opened since end of 
WGLC (working group last call) which are related to this analysis:

<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=76>: "What event 
does creationdate refer to" -- I'd like not to say anything particular 
about DAV:creationdate because that would imply it's different from 
other WebDAV properties. However, *if* we say something, at least it 
needs to be both correct and non-misleading (thus we'll need to state 
things about other WebDAV properties as well).

<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=77>: "Do Bind 
operations change the getlastmodified or getetag property values" -- 
Yes, they may. See above. BIND can't IMHO make any new requirements 
here; thus all we could say is that RFC2616 applies (which it does even 
if we don't say so).

<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=78>: Value of 
ETag and getlastmodified properties on multiple bindings> -- same as above.

I'd also like to ask the WG chairs to close the other open issues 
related to BIND 
(<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=2>, 
<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=5>, 
<http://ietf.cse.ucsc.edu:8080/bugzilla/show_bug.cgi?id=71>) as agreed 
upon in Last Call process 
(<http://lists.w3.org/Archives/Public/w3c-dist-auth/2005JanMar/0001.html>).


Best regards,

Julian