- From: Roy T. Fielding <fielding@kiwi.ics.uci.edu>
- Date: Wed, 30 Sep 1998 16:57:41 -0700
- To: "Babich, Alan" <ABabich@filenet.com>
- cc: WebDAV WG <w3c-dist-auth@w3.org>
There are a few things that need to be understood about WebDAV resources
and variants.
o All variants are resources. Whether or not they are accessible
resources is a separate question. Whether they are "stored" or
"derived" is irrelevent to their nature as variants -- they behave
no differently than other resources that are "stored" or "derived".
o Resources that serve as negotiation handles are, by definition,
derived resources. They can be authored in the same way as any
CGI output or SSI output, or simply redirected with a 300 response.
o WebDAV identifies resources; it does not have any say over how they
are implemented. A server implementation does need to be aware of
how its resources are implemented so that it can determine how to
handle a resource-modifying request, but that distinction can be
boiled down to "stored" or "derived", regardless of the resources'
other properties or their relation to other resources.
In summary, one goal of WebDAV versioning might be to support the
versioning of derived resources. This is essentially a configuration
problem, since derived resources are composed of other resource(s)
which are also versionable. Personally, I consider this to be just
another form of collection graph.
OTOH, that goal may be a bit hard to accomplish. I don't know, since
it really depends on how quickly everyone can agree to a general
interface model upon which to base a protocol.
....Roy
Received on Wednesday, 30 September 1998 20:02:58 UTC