RE: bind draft issues

Actually, my suggestion below that RFC3253 is neutral wrt to
the meaning of "resource identity" is probably somewhat misleading.
In particular, there is the underlying assumption that you can
make any version of a version-controlled resource appear at the
location of that version-controlled resource (e.g. with an UPDATE
request).  This assumption is not very compatible with an attempt
to have a URL define the "variant" of a resource (e.g. the Swedish
variant), since any version checked-in at the Swedish variant URL
would go into the version history and then be accessible at any
other version-controlled resource for that version history.
In addition, wrt to PROPFIND semantics, many important use cases
would be broken if PROPFIND of a live versioning property could
return different values at different URL locations.

So there is a basic leaning towards "content and properties are the
same at all URL locations of a given resource" semantics for a
versioning server (at least, for version-controlled resources).

Cheers,
Geoff


-----Original Message-----
From: Clemm, Geoff [mailto:gclemm@Rational.Com]
Sent: Monday, March 10, 2003 9:09 AM
To: WebDAV; Roy T. Fielding
Subject: RE: bind draft issues



   From: Julian Reschke [mailto:julian.reschke@gmx.de]

   Geoff, as RFC3253 relies on the concept of bindings, could you
   please attempt to fill in what *you* think an identical
   DAV:resource-id needs to indicate?

For RFC3253, two version-controlled resources are the "same" when
they have the same version history.  So the obvious implementation
of the DAV:resource-id property for a version-controlled resources
is the DAV:version-history property.

Version and Version-History resources cannot be moved, so the
URL at which the were created can act as their DAV:resource-id.

So this allows RFC3253 to remain neutral on the question of 
what does "resource identity" mean about the results of various
other HTTP and WebDAV methods.

Note: the question of resource identity for an activity resource
is not addressed by RFC3253, so that remains an open question.
Since most versioning repositories store activity resources in some
kind of database, resource identity for activities ends up not
being much of a problem in practice though, since there usually will
be some obvious implementation property that could be used as
the DAV:resource-id property.

Cheers,
Geoff

Received on Monday, 10 March 2003 09:24:24 UTC