W3C home > Mailing lists > Public > ietf-dav-versioning@w3.org > January to March 2001

RE: Autoversion confusion

From: Clemm, Geoff <gclemm@rational.com>
Date: Fri, 9 Feb 2001 14:39:45 -0500
Message-ID: <3906C56A7BD1F54593344C05BD1374B1018E2239@SUS-MA1IT01>
To: ietf-dav-versioning@w3.org
I'd be happy to see this language in the FAQ
(and strongly encourage Lisa to add it there),
but I think it would be confusing to put something
in the protocol that in effect says "for this resource,
GET and PUT will continue to work as defined in HTTP",
since a natural conclusion would be that we are *not*
guaranteeing that they will do so in other cases.
And we definitely don't want people to reach that
(incorrect) conclusion.

Cheers,
Geoff

-----Original Message-----
From: Tim_Ellison@uk.ibm.com [mailto:Tim_Ellison@uk.ibm.com]
Sent: Friday, February 09, 2001 11:47 AM
To: ietf-dav-versioning@w3.org
Subject: RE: Autoversion confusion




Lisa wrote:
> I wasn't talking only about the client that did
> the put.

Neither was I.  In general the HTTP sever has no way of knowing which
client is sending a PUT/GET request.

> The DeltaV spec ought to state normatively what
> OTHER clients experience.

Thankfully the spec doesn't get into sessions or whatever for identifying
clients.  It's unnecessary.

> I'll repeat from my earlier mail, my suggested language:
>
> "In core versioning, while a VCR is checked out it
> may be the target of multiple write operations.  During
> this period, other clients MUST still be able to perform
> read operations on the VCR's URL, and the results MUST
> show the results of all the write operations that have
> been performed thus far.  (Note: if a scenario requires
> hiding a work-in-progress from other clients, the "working
> resource" option can be used.)"

I'd be interested to hear if others think that we need to state this.


Tim

------------------------
> > -----Original Message-----
> > From: ietf-dav-versioning-request@w3.org
> > [mailto:ietf-dav-versioning-request@w3.org]On Behalf Of
> > Tim_Ellison@uk.ibm.com
> > Sent: Friday, February 09, 2001 1:58 AM
> > To: ietf-dav-versioning@w3.org
> > Subject: RE: Autoversion confusion
> >
> >
> >
> >
> > John wrote:
> > > This shouldn't be necessary, since the HTTP spec
> > > defines the behavior of GET and PUT. Specifically,
> > > it says that PUT to a particular resource defines
> > > the response for any following GET on that same
> > > resource (I'm paraphrasing from memory). There
> > > can't be any other possible interpretation (that
> > > doesn't break HTTP semantics).
> >
> > I agree with John.
> >
> > At the risk of nagging<g>, a version-controlled resource is
> > an honest to
> > goodness WebDAV resource, with content and properties
> > (version-controlled
> > collections have members and properties).  Intuatively, if a
> > PUT to the
> > resource succeeds (200 OK) then a client is entitled to
> > believe they will
> > GET the same entity back.
> >
> > p.s.
> > I had a quick look through the HTTP/1.1 spec, and didn't see
> > anything that
> > states this categorically.  In fact, Section 9.6 (PUT) states:
> >  "HTTP/1.1 does not define how a PUT method affects the state
> > of an origin
> > server."
> > Now, how many clients do a GET just to check what they actually will
> > retrieve after a successful PUT!
> >
> > Tim
> >
Received on Friday, 9 February 2001 14:31:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 13:57:40 GMT