RE: WebDAV Bindings - Issue Yaron.ApplePieToo

I agree that we want to keep the sentence you identified below.

I also agree that we want to address the issue of a resource acting
differently based on the URI it is addressed from.

I suspect my problem with the language in the section derives from the
differentiation made between static and dynamic resources. In my experience
one can not really make the distinction because even if a "resource" is just
a file in a file system, there is always a server on top of it that may or
may not alter its presentation. If the language in section 9 were to be
re-written with the assumption that ALL resources are dynamic then I suspect
we would find the resulting language mutually acceptable. 

In other words, we should remove the terms dynamic/static completely and
instead talk about resources and the fact that their presentation can change
based solely on what URI they are accessed from. We shouldn't even try to
address the fact that in some cases, some resources for some period of time
may choose to look identical regardless of what URI they are accessed from.

		Yaron

> -----Original Message-----
> From: Slein, Judith A [mailto:JSlein@crt.xerox.com]
> Sent: Tue, January 18, 2000 7:31 AM
> To: 'Yaron Goland'; w3c-dist-auth@w3.org
> Subject: RE: WebDAV Bindings - Issue Yaron.ApplePieToo
> 
> 
> The first paragraph doesn't really say anything substantive.  It just
> introduces the section, listing the methods to which it 
> applies.  The one
> substantive statement that I think we would want to keep is "For these
> methods, no new complexities are introduced by allowing 
> explicit creation of
> multiple bindings to the same resource."
>  
> I could be persuaded to reduce the section to just that.
>  
> But it still might be worth pointing out that BIND will make 
> it much more
> common to have multiple bindings to the same resource (even 
> though that was
> always possible).  And when there are multiple bindings to a 
> resource, it is
> possible for a resource's behavior in response to GET, HEAD, 
> PROPFIND, etc.,
> to be sensitive to the URI that was used to make the request. 
>  Maybe we
> wouldn't have to discuss the distinction between static and dynamic
> resources in order to make that point.
>  
> --Judy
>  
> 
> -----Original Message-----
> From: Yaron Goland [mailto:yarong@Exchange.Microsoft.com]
> Sent: Sunday, January 16, 2000 8:28 PM
> To: w3c-dist-auth@w3.org
> Subject: WebDAV Bindings - Issue Yaron.ApplePieToo
> 
> 
> 
> Section 9 of the draft attempts to tackle the problem of how to
> differentiate the behavior of a resource based on its 
> bindings. It attempts
> to discuss the difference between a static and non-static 
> resource. However
> the difference between these two types of resources is so unbelievably
> arbitrary that I fail to see how attempting to differentiate 
> between them
> enhances interoperability.
> 
> For example, let us pretend we have the CurrentURI property. 
> The value of
> the property will be the Request-URI of the PROPFIND method 
> used to retrieve
> the value of the CurrentURI property. Now let us pretend we have two
> servers, A and B. A runs a version of WebDAV that 
> automatically generates
> the CurrentURI property on top of whatever resources it 
> supports. What this
> means is that even if the underlying data store where the resource is
> recorded doesn't know about the CurrentURI property, I can 
> still get the
> CurrentURI property through PROPFIND because A's server will 
> intercept the
> PROPFIND and insert the data itself. Server B doesn't support 
> CurrentURI at
> all. I now have a resource R that has bindings into A and B. 
> This means that
> if I ask for the CurrentURI property through the binding on 
> A, I will get a
> result with the URI on A. If I ask for the CurrentURI 
> property through the
> binding on B, I will not even see the property at all.
> 
> So the question is - Well is this a dynamic resource? Is it 
> not? Can I have
> the same resourceID on both URIs in this example? How does 
> the language in
> section 9 help me deal with this? Does it actually clarify 
> anything in a
> useful way?
> 
> As far as I can tell the language in section 9 is just 
> motherhood and apple
> pie language and that always makes for bad standard language.
> 
> Therefore I move that all the paragraphs in section 9 but the first
> paragraph be stricken from the BIND spec. 
>

Received on Tuesday, 18 January 2000 11:13:22 UTC