W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2006

Re: Etag-on-write, 2nd attempt (== IETF draft 01)

From: Julian Reschke <julian.reschke@gmx.de>
Date: Tue, 12 Sep 2006 20:10:52 +0200
Message-ID: <4506F82C.2010304@gmx.de>
To: Joe Gregorio <joe@bitworking.org>
CC: HTTP Working Group <ietf-http-wg@w3.org>

Joe Gregorio schrieb:
> Editorial nits:
> 1. " programmers asked servers to always return "ETag" headers upon
>   PUT, never ever to change t"
>   I can't follow that sentence. Sorry, no construtive suggestions on my 
> part.

Any kind of feedback is useful :-) In this case, it's a typo, it's 
supposed to say:

"...client programmers asked server programmers..."

> 2. Section 1.2
>    "reasons for HTTP's success, allowing to be used for a wide range of"
>     allowing IT to be


> General:
> How is "Entity-Transform: identity" any different than returning
> a strong ETag as opposed to a weak ETag?
>   -joe

The issue is that without further information, a client can not assume 
that upon PUT, the entity was stored octet-by-octet, thus it needs to 
refetch the content if it needs that kind of equivalence. This may not 
be obvious from RFC2616, but it is certainly the conclusion we arrived 
at over here on the HTTP mailing list something like 10 months ago. Thus 
the proposed clarifications in Section 3, and the "Entity-Transform" 
extension header definition.

As this is the key point of the whole exercise, I would really 
appreciate if you could re-read the draft with an eye on this, and let 
me know whether you agree with the analysis of the problem in the first 
place :-)

Thanks a lot,

Received on Tuesday, 12 September 2006 18:40:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 11:10:40 UTC