RE: again: OPTIONS semantics

Yes, I think it would be fine to consider different backends to be
different servers.  The only thing I'd insist on is that the OPTIONS
request return "update" in the DAV header if the resource identified
by the request-URL supports the UPDATE method.  Beyond that, I'd
just do whatever makes sense for your implementation.

Cheers,
Geoff


-----Original Message-----
From: Julian Reschke [mailto:julian.reschke@greenbytes.de]

let's take that further -- this is a real implementation issue, not just
academic.

I have *one* servlet "/a" that controls access to different backends through
different namespace partitions such as "/a/filesystem", "/a/jdbc" and so
on... One backend may support workspaces, the other won't.

Would you still consider this "multiple" servers?
--
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760 
-----Original Message-----
From: ietf-dav-versioning-request@w3.org
[mailto:ietf-dav-versioning-request@w3.org]On Behalf Of Clemm, Geoff
Sent: Thursday, September 26, 2002 11:22 PM
To: ietf-dav-versioning@w3.org
Subject: RE: again: OPTIONS semantics


I would classify two different servlets as two different 
servers, and so if the servlet handling the OPTIONS request 
does not know about the update feature, then it is expected 
to not return "update" in the DAV header. 
Cheers, 
Geoff 
-----Original Message----- 
From: Julian Reschke [mailto:julian.reschke@gmx.de] 
Sent: Thursday, September 26, 2002 5:16 PM 
To: ietf-dav-versioning@w3.org 
Subject: again: OPTIONS semantics 



Just taking an example from section 7: 
"Additional OPTIONS Semantics 
If the server supports the update feature, it MUST include "update" as a 
field in the DAV response header from an OPTIONS request on any resource 
that supports any versioning properties, reports, or methods." 


So assuming my DAV namespace is separated into two parts (for instance by 
servlet) "/a" and "/b". Both parts of the server namespace support 
versioning, but only "/a" supports workspaces. Should the "update" feature 
be reported on "/b" and it's descendants? 
The spec says "yes", but I don't think this is implementable (for instance, 
the servlet responsible for "/b" may not even know about the existence of 
"/a"). 
Julian 
-- 
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760