Re: What is a specification for? [was Re: Calculating Age Question]

On Wed, 27 Nov 1996, Roy T. Fielding wrote:

> As an author of the specification, I have a right to point out those
> parts of the specification which are known to be in error.  It is foolish
> to insist that implementers should implement HTTP/1.1 incorrectly just
> because the specification was accepted by the IESG with a known error
> inside of it.  This is not a matter of opinion, it is a fact based on
> both the original intent of the Age header field and the intent of the
> caching subgroup.  I have provided sufficient documentation of that fact,
> and your opinion that the age calculation should be "conservative" is
> irrelevant to whether or not the Age header field should be created by
> a proxy that never used its cache in forwarding the response message.

Give it up Roy. You have been trying to push your personal belief on this
for months. The most that can be *accurately* said is that you object to
the clear statement in the specification. You have no special 'rights' -
the spec is the *consensus* of the HTTP WG, not your personal spec.  That
you do not *LIKE* the final spec is not a problem with the spec but with
your relationship to it.

> 
> Encouraging people to implement Age as you suggest would promulgate
> incorrect implementations and make Age useless as a real lower-bound
> for the already-conservative age calculation.  If it becomes useless,
> then you have defeated its purpose entirely.  I cannot wait until the
> next round of drafts if I am to prevent Age from becoming useless,
> nor do I have the time to generate another special draft just to point
> out an error which should be obvious to anyone who actually looks at
> the problems caused by promulgating erroneous Age values.

Roy - false cachable hits are MUCH worse than false uncachable hits. Your
'doomsday scenario' of network meltdown because of the extra hits due to
pre-maturely expired objects has NOT been demonstrated and some people
pointed out that your prefered scheme potentially results in stale objects
being served as fresh - which is utterly unacceptable under any
circumstances. Caches *should* be ultra-conservative with regard to Age.

The published proposed standard is correct. Your position is not. Please
quit trying to express an ownership of the standard - it is not "Roy
Fielding's Proposed HTTP Standard". It is the IETF's proposed standard
and the result of the consenus of the WG. 

-- 
Benjamin Franz

Received on Thursday, 28 November 1996 07:41:55 UTC