Re: Etag-on-write, 4rd attempt (== IETF draft 03), was: I-D ACTION:draft-reschke-http-etag-on-write-03.txt

   And to clarify, so long as Julian's formula holds true:

	t(e2(e1(c))) = t(e2(t(e1(c))))

   A client doesn't necessarily need to care whether an ETag is strong  
or weak.

   CalDAV is a prime example of where weak ETags are appropriate.  If  
the underlying data store isn't iCalendar, and the server needs to  
render iCalendar data every time a resource is requested, it's  
entirely possible that the iCalendar data is semantically but not  
binary equal every time.  Yes, you can make sure your iCalendar  
generation is deterministic, but if you change the software, you need  
to change the strong ETags for every resource on your server along  
with it.  That's inane.  CalDAV dropped the ball here.

	-wsv


On Nov 5, 2006, at 5:54 PM, Wilfredo Sánchez Vega wrote:

>   Well, that's going to be dependent on the delivered format.  Many  
> formats have multiple representations which are considered  
> equivalent.  This includes iCalendar, which is why CalDAV's shoo- 
> shoo-go-away stance on weak ETags is, well, weak.
>
>   The example I gave, where the order of iCalendar properties  
> changes within an iCalendar component, is what I have in mind in  
> that context.
>
> 	-wsv
>
>
> On Nov 5, 2006, at 1:29 PM, Lisa Dusseault wrote:
>
>> Speaking for myself, my aversion is to ambiguously-defined weak  
>> ETags.  You could probably define something called "Weak ETags" and  
>> say more about how they work than RFC2616 does, and create  
>> something useful.  I'm sure your idea of what weak ETags do is a  
>> sane one and if everybody agreed we'd have interoperability.  But  
>> we've had interoperability problems around weak ETags and that's  
>> the root cause of my aversion.
>>
>> Lisa
>>
>> On Nov 4, 2006, at 1:33 PM, Wilfredo Sánchez Vega wrote:
>>
>>>   As a general note, I still don't quite understand the widespread  
>>> aversion to weak ETags, and I think that the this editing property  
>>> show why weak ETags *do* work in an authoring environment.
>

Received on Monday, 6 November 2006 19:19:17 UTC