Re: Content Type

> Any classification scheme could be used, but I still think percentages are
> useful as a compromise between specifying each and every structure
> supported and pretending support for "level A" when it in fact supports
> only 80% of it.
>   
This is precisely the problem IMO, there is no way to tell what part 
those 80% really cover. Firefox 80% could be different from Opera 80% 
and we are still in trouble.
> "Atomic" negotiation would be useful, but I believe it would have to be
> moved to a separate request/response to avoid information overload.
>   
Hmmm that sounds fair although I wonder if it would not make it more 
complex.

- Sylvain

Received on Wednesday, 22 March 2006 10:52:14 UTC