RE: GET, POST, and side-effects

Paul Leach writes:
 > It seems pretty straightforward that if a proxy forwards a POST, and 
 > gets back a Location: URI and Expires: or Cache-Control: headers that 
 > say that the returned entity will be fresh for some period of time, it 
 > should be able to serve that entity safely from the cache for that 
 > period of time,

for subsequent GETs, yes.  For subsequent POSTs, not in general (because of side
effects on the origin server).

 and do GET I-M-Ss on it after that time to determine if 
 > it is still fresh.  Maybe very few POSTs will do this, because most 
 > POSTs are for dynamically generated content, but that doesn't affect 
 > the correctness of the argument.

Dynamically generated content can be cached.  It's just hard to take
advantage of that without a way for non-side-effecting POSTs to 
be serviced from a cache.

 > 
 > A very similar argument applies to PUT.
 > 
 > Paul
 > 
 > 

Shel

Received on Thursday, 4 January 1996 22:44:02 UTC