On Jan 5, 2008, at 01:10, Bertrand Le Roy wrote:

> I still think it would be interesting to investigate precedents as  
> this seems at first to deviate from usual HTTP GET usage.

It is very clear that the spec deviates from usual HTTP GET usage. The  
HTTPish way would be using OPTIONS with a new response header that had  
application-level caching semantics.

However, OPTIONS has been rejected due to issues in the popular Apache  
server with certain modules.

Henri Sivonen

Received on Saturday, 5 January 2008 11:04:17 UTC