durusmail: quixote-users: proposition for a better Browser caching mechanism
proposition for a better Browser caching mechanism
2005-05-15
2005-05-17
2005-05-17
2005-05-19
2005-05-22
proposition for a better Browser caching mechanism
Graham Fawcett
2005-05-19
On 5/17/05, vincent delft  wrote:
> --- Neil Schemenauer  wrote:
> > Not all Quixote servers speak HTTP/1.1.  I think we
> > should stick to
> > HTTP/1.0 if possible (the Host header is one
> > exception).
>
> I'm sorry, I don't catch your point. Webservers send
> headers and content to browsers. I'm just asking to
> replace the "expires" by "cache-control".

I think Neil's point is that "Web servers send headers and content to
browsers" according to *specifications*. There are two of versions of
the spec in the real world: HTTP/1.0 and HTTP/1.1. The former
specification does not specify "Cache-control" or most other caching
directives. To be compliant, you cannot send a 1.1 header from a 1.0
server. Postel's Law and all of that. (But Neil threw "Host" in there
as a 1.0 retrofit, so we're already muddying the waters!)

I reckon that Quixote could be smart enough to always do the right
thing, if told via configuration what version of HTTP was in effect.

Wondering-how-many-http/0.9-servers-are-still-out-there'ly yours,
-- Graham
reply