On Tue, 12 Dec 2006 23:25:44 +0100, Tim Bray <[EMAIL PROTECTED]> wrote:
[...] So I see no downside in James doing an I-D.
But is a separate I-D really necessary? If, like Kyle Marvin suggests, the new MIME type for Atom Entries actually becomes a type parameter of the existing Atom MIME type, can't the language that specifies this be included in the APP specification?
By that, APP will only extend RFC 4287 and not deprecate anything. It might include wording like "SHOULD use the type parameter", but that still makes the old and bare MIME type valid. Thus, implementing RFC 4287 will not require you to use a type parameter, but implementing the APP specification will strongly encourage you to do so. Won't that solve it all?
-- Asbjørn Ulsberg -=|=- http://virtuelvis.com/quark/ «He's a loathsome offensive brute, yet I can't look away»
