Re: proposed HTTP changes for charset

On Mon, 8 Jul 1996, Shel Kaphan wrote:

> Benjamin Franz writes:
> 	...
>  > 
>  > Ok. All of these cases work ok. So the problem has got to be when you
>  > stick a proxy in the line. How does mandating charsets break proxies?
>  > I don't see it.
>  > 
>  > -- 
>  > Benjamin Franz
>  > 
>  > 
> 
> One case is when a 1.1 proxy receives a document from a 1.0 server,
> and it is unlabelled.  The proxy stores the document in its cache, and
> on a later request from a 1.1 client, has to do something about the
> charset.  If charset labelling is mandatory the proxy has to guess,
> which is not going to work.  So if charset labelling is mandatory in
> 1.1, either the proxy has to have some way of indicating the content
> has an unknown charset, or (ugh) it would have to revert to 1.0
> protocol so that it could legally send an unlabelled response.

Reverting to 1.0 may not be pretty - but it has the tremendous virtue of
*working*. It seems the right thing to do in any case. Attempting to
'upgrade' a response from 1.0 to 1.1 seems questionable practice at best
and promises to break things.

-- 
Benjamin Franz