Bill de h�ra wrote:
If,

 - 6.2 was dropped and probably
 - the first sentence of the Pace was dropped,
 - the rest of the 1st para was dropped down into 6.1
 - there was some weasel wording about RFC3023 compliance

how would you feel about the rest of it?
...

I think the main issue here is first we say

1) ...SHOULD use application/atom+xml...

...then, if you don't want to...

2) ...SHOULD use application/xml...

...then finally...

3) otherwise MAY use text/xml.

That is a problem in itself. Either we mandate a specific content type or we don't (I think we should). If we mandate one, we should say that behaviour of documents served with a different type is simply undefined.

Note that most of the nasty RFC3023 problems only apply to text/*, in particular I don't see why we would want to RECOMMEND to use a charset parameter on application/* content types.


Best regards, Julian


-- <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760



Reply via email to