Re: Caching Servers Considered Harmful (was: Re: Finger URL)

Daniel W. Connolly (connolly@hal.com)
Mon, 22 Aug 1994 13:08:54 -0500


In message <Pine.3.89.9408221253.B1629-0100000@sgf.fv.com>, Darren New writes:
>
>Well, it certainly sounds like the protocol needs a time-to-live, now
>doesn't it?

AAAAAAAH! See what you've done, Mr. Raisch?? Another specification
issue that was wrestled to the ground long ago, all muddled up again.
You should know better.

The protocol HAS a time-to-live!

>From "Object Header lines in HTTP" at
http://info.cern.ch/hypertext/WWW/Protocols/HTTP/Object_Headers.html#z11

Expires: date

Gives the date after which the information given ceases to be
valid and should be retrieved again. This allows control of
caching mechanisms, and also allows for the periodic
refreshing of displays of volatile data. Format as for
Date:. This does NOT imply that the original object will cease
to exist.

Everyone: right now: before you post another message to www-talk,

READ THE SPECS!

They're all online, reachable from:

http://info.cern.ch/hypertext/WWW/Technical.html

If you don't like them, get them fixed!

<rant>

It's no wonder the WWW specs are so poorly maintained. You not only
have to design, write, and edit the dang thing, but because folks
don't actually read the specs (partly because they're out of date
because nobody wants to edit them...), you have to watch the
discussion to be sure that implmentors don't get the wrong idea and
implement something else, because we all know that the specs don't
really matter... the existing implementations are what really
matter...

AAAARGH! Do we really expect to build a large, complex, distributed
hypermedia system with such lack of diligence?!?!?

</rant>

Dan