I agree with Julian that there is a problem with requiring support for an unspecified normative feature. It is better to reformulate in informative language that recommends sniffing without using any conformance keyword. On Mon, Sep 8, 2014 at 4:32 AM, Robin Berjon <robin@w3.org> wrote: > On 08/09/2014 12:19 , Julian Reschke wrote: > >> On 2014-09-08 12:03, Robin Berjon wrote: >> >>> I hear your discomfort, but I think the above would (meaning no >>> disrespect) give lawyering precedence over pragmatism. Sniffing is not >>> optional; it is, however, defined orthogonally. >>> >> >> How can it be non-optional if you don't have a spec defining it? >> > > Because for better or for worse we live in a world that is not entirely > defined by specifications, but also in decreasing but still important part > by existing deployed implementations. It's not optional in that you really > just can't write a (useful) browser without this. > > The fact that there isn't a spec is a problem (in general, for the Web) > but isn't a specific concern of HTML. > > How about this: I add a warning to the reference that conveys your >>> concerns? >>> >> >> Yes, please make a proposal! >> > > """ > As of today the Web community lacks a sufficiently complete, reliable, > interoperable, and tested specification for the manner in which content > sniffing takes place on the Web. We encourage implementers to exercise > caution in this area as the Web community makes progress towards addressing > this issue. > > """ > > -- > Robin Berjon - http://berjon.com/ - @robinberjon > >
Received on Monday, 8 September 2014 10:37:28 UTC