W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2001

RE: Issues/questions about the bindings protocol specification...

From: Clemm, Geoff <gclemm@rational.com>
Date: Thu, 13 Dec 2001 02:29:02 -0500
Message-ID: <3906C56A7BD1F54593344C05BD1374B1052AD4AA@SUS-MA1IT01>
To: w3c-dist-auth@w3.org
   From: Peter Raymond [mailto:Peter.Raymond@merant.com]

   1) Section 9 says that methods MUST produce the same results when
      submitted on different bindings to the same resource.  I
      disagree, if you issue a PROPFIND request on /coll1/bar the href
      returned would be /coll1/bar, if you issue the same PROPFIND
      request on /coll2/bar (which is bound to the same resource) the
      href returned would be /coll2/bar.  The responses would not be
      the same.  I am sure there could be other examples of this.

Yeah, trying to define the semantics of a "binding" is tricky.
I'd have to look up in our notes whether we came up with something
better, or just didn't try.  We might just need to give a method
by method definition of the semantics, and then let folks infer
from that how they should act for future extensions.

   2) We found the definition of the DAV:bindings property in section
      13.1 very confusing.  It would be good to show an example of
      this property, an interesting example would be the DAV:bindings
      on resource R2 from Figure 1.  I will send a separate e-mail on
      this issue.

OK.

   3) Section 15 describes capability discovery and shows the use of a
      header called Public.  I cannot find a definition of this header
      in RFC2518 or RFC2616 (there is a Cache-Control public directive
      but no Public header).  In what specification is this header
      defined?  Maybe I just missed it.

This is a deprecated header from an earlier version of the HTTP-1.1
draft (rfc 2068).  The reference should be removed.

   4) Also on the subject of section 15, it seems we have many ways of
      finding out what capabilities a server or a resource has...Allow
      header, Public header, DAV header, DAV:supported-method-set
      property (in DeltaV).  I find this confusing, and I know in the
      various WebDAV groups there has been some discussion of these
      and the differences between them, I think they should be defined
      clearly and consistently in all the WebDAV related
      specifications.

The Public header is gone, but the rest remain.  We have defined 
DAV:supported-method-set to be the same as Allow, so it really
is DAV header, combined with the DAV:supported-xxx properties.
And we hope to add these properties to the next 2518 draft, so
that they are generally available, and not just in DeltaV.

   5) Section 18 there is a missing space in the last sentence (after
      the 507 code).

Yup.

   6) Is there anyone looking at how bindings affect the other
      protocols, for example have we looked at how DeltaV features
      (like Baseline etc) are affected by bindings?

I tried to keep bindings in mind for all of the of the DeltaV
features.  If you notice anything that might be a problem, please let
me know.

   BTW...are there any server vendors out there that implement or are
   thinking of implementing the BIND method?

Yes.

  Similarly are any clients using it? 

Not yet, as far as I know.

Cheers,
Geoff
Received on Thursday, 13 December 2001 02:29:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:59 GMT