Re: ETags?

And just to be clear (and to apologize for being a bit testy
in the message below :-), Elias was not the one posting requests
for implementation help ... his posts have all been very reasonable (:-).

Cheers,
Geoff

Geoff wrote on 01/19/2005 10:24:25 PM:
> 
> If there were a well-defined (and finite) number of such "guidances", 
> I could live with it (and it was on that basis that I reluctantly 
> agreed to adding the "guidance" text about locking). 
> 
> But as soon as we agree to one such "guidance", a new one is suggested. 
> I despair of ever getting the BIND protocol published if it is delayed 
> until it has become a complete "implementation guide". 
> 
> I noticed that yet another such request for "guidance text" has just
> been posted to the 
> bug database [bug 71].  I also note that nothing in the "bug" has to 
> do with any of the methods or properties defined by the binding 
specification,
> but is about access control. 
> 
> I have nothing against asking for guidance from the mailing list, 
> but holding a specification hostage until all of your implementation 
> issues have been addressed by the mailing list seems a bit unreasonable. 
 
> 
> Cheers, 
> Geoff 
> 
> 
> Elias wrote on 01/19/2005 09:53:43 PM:
> >
> > Geoffrey M Clemm wrote:
> > 
> > > I agree with Roy's rationale and conclusion, and support the removal 

> > > of the reference to live properties in section 2.6
> > 
> > So, in effect, the spec will state that dead properties MUST be the 
same 
> > across all bindings to a given resource, and that (by remaining silent 

> > on the issue) live properties MAY vary depending on server 
> > implementation. I am not oposed to this, however as a guide to client 
> > implementors it might be nice to point out that they can't 
neccessarily 
> > depend on live properties being the same across all bindings to a 
resource.