Here's my proposal: * Replace the first line of Part 6, section 3.1: - A correct cache MUST respond to a request - with the most up-to-date + For methods other than GET and HEAD, a cache + MUST forward the request to the origin server + and MUST return the origin server's response. + For GET and HEAD requests, a cache MUST + response with the most up-to-date * Insert the following paragraph at the beginning of part 6, section 6 (Response Cacheability): + Responses for methods other than HEAD and GET + MUST NOT be cached. * Remove part 6, section 13 (Write-Through Mandatory). * Remove all the "responses to this method are not cacheable" statements from part 2. * Remove the statement about the cacheability of POST in part 2, section 8.5. * Issue an errata or some other update of RFC 4918 that says that PROPFIND responses are not cacheable. Justification: * This is the simplest thing that could possibly work. * Nobody is caching anything other than GET and HEAD responses anyway. * Section 13 is redundant with the other parts, and it has misleading statements. For example, according to section 13, a cache cannot return a 5xx response if it cannot communicate with the server. - BrianReceived on Tuesday, 18 November 2008 19:50:26 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:34 GMT