RE: whenToUseGet-7 counter-proposal

> I would suggest clarifying this by adding the sentence:
> 
> "Systems should NOT however, rely on HTTP GET to be 'safe'"

The expected behavior of HTTP GET is a different (albeit
admittedly related) topic than "When to use GET".

RFC 2616 section 9.1 "Safe and Idempotent Methods" already has
a discussion of HTTP GET and safety. I'd hate to see a
TAG recommendation on the topic that didn't at least reference
the HTTP spec.

There's been some talk of an update to RFC 2616, at a minimum
to incorporate the errata (http://purl.org/net/http-errata).

Received on Wednesday, 24 April 2002 11:48:23 UTC