W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2004

Re: [Bug 3] Bindings draft should specify if all properties MUST have same value on all bindings

From: Lisa Dusseault <lisa@osafoundation.org>
Date: Mon, 6 Dec 2004 14:01:37 -0800
Message-Id: <666E83CC-47D2-11D9-A94C-000A95B2BB72@osafoundation.org>
Cc: Webdav WG <w3c-dist-auth@w3c.org>
To: Julian Reschke <julian.reschke@gmx.de>


On Dec 6, 2004, at 12:55 PM, Julian Reschke wrote:

> Lisa Dusseault wrote:
>
>> Only if RFC2518bis is going to include the whole bindings 
>> specification, or at least the resource-id property.  The reason why 
>> RFC2518 never needed to answer this question definitively is because  
>> clients had no way of knowing when two URLs point to the same 
>> resource.  If clients can't identify bindings, then there's no way to 
>> violate the client's expectations around property values of two 
>> bindings to the same resource.
>
> If it doesn't matter, why does RFC2518 say:
>
> "Although implicit in [RFC2068] and [RFC2396], any resource, including 
> collection resources, MAY be identified by more than one URI. For 
> example, a resource could be identified by multiple HTTP URLs."
>
> (<http://greenbytes.de/tech/webdav/rfc2518.html#rfc.section.5.1.p.4>) 
> and
>
> "A resource may be made available through more than one URI. However 
> locks apply to resources, not URIs. Therefore a LOCK request on a 
> resource MUST NOT succeed if can not be honored by all the URIs 
> through which the resource is addressable."
>

In this case it *does* matter to clients because it is detectable -- 
the behavior has concrete results, even if the client can't detect that 
there are multiple bindings.  If a client is trying to allow the user 
to edit a resource and therefore locks the resource, then the resource 
gets changed anyway, this would violate the client's expectation.  The 
client probably doesn't have a graceful way to recover from this kind 
of event even if the programmers are responsible in checking for it (by 
looking at the ETag).  Therefore, this language is necessary to ensure 
that the lock covers the real resource and that the lock behaves the 
way the client expects it to.

If we can come up with a use case where it's important for the client 
to rely on property values being the same for every binding to a 
resource, even without being able to tell that the resource has 
multiple bindings or what they are, then I will agree that RFC2518bis 
needs to cover that behavior as well.

Lisa
Received on Monday, 6 December 2004 22:01:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:17:51 UTC