- From: David W. Morris <dwm@shell.portal.com>
- Date: Thu, 4 Jan 1996 21:52:28 -0800 (PST)
- To: HTTP Caching Subgroup <http-caching@pa.dec.com>
On Thu, 4 Jan 1996, Paul Leach wrote: > Note, finally, that it is not necessary for the same query to give the > same result _forever_ -- only that it give the same result for as long > as the Expires: and/or Cache-Control: headers say proxies are allowed > to believe it does. It is not necessary even for a GET to return the same result, only that the action of making the GET request won't change the result or have other side-effects. Any large constantly changing document repository can be expected to respond with a different result as document are added. BUt the GET doesn't cause this effect. The cache-control mechanisms allow such a server to declare how long the server finds it acceptable to provide the same response to the GET no matter what the server would deliver itself. 'freshness' is a judgement of the information provider which may choose to define search results as static for 24 hours even though in fact they vary by the second. Other applications will use the same cache control mechanisms to forbid caching because the risk associated with using inexact data is too great. Dave Morris
Received on Friday, 5 January 1996 06:02:27 UTC