Re: Versioning and html[5] : application/xhtml+xml

Chris Wilson schrieb:
> Dão Gottwald [] wrote:
>> Alan Dean schrieb:
>>> 1) IE7 will not render an xhtml document presented as
>>> application/xhtml+xml.
>>> 2) There is, therefore, nothing to break (it already doesn't work).
>>> 3) Given 1 and 2 above, is it feasible to support a 'clean' html[5]
>>> when presented as application/xhtml+xml and preserve backwards
>>> compatibility when presented with text/html?
>> Roughly the same question can be (and was) asked for <!DOCTYPE html>.
>> (It is supported by IE7, but not used in today's documents.) Ideally, I
> That's not exactly the same, though it's close.  The major difference is that with the different MIME type, IE will do nothing with it at all - pop a "save as" dialog, I think - so it is not possible to use it (unless you configure your server to send it as text/html).  You can use <!DOCTYPE html> today - although virtually no one does today.

Yep, that's what I meant.

>> want IE to have one decent rendering mode for HTML5 and above,
>> regardless of the mime type.
> Noted.
>> If Microsoft supports such a mode before
>> the significant bugs are fixed and without an opt-in, it'll be just
>> another quirks mode and we need yet another doctype or mime-type switch.
> Who defines "significant"?

You, Microsoft. As usual, outsiders like me can make suggestions, and 
I'll be happy to do just that, in case you release beta versions. But 
it's up to you to draw the line, as it's up to you to define how many 
broken sites you can afford if you update the new standards mode once 
it's deployed.

You've recently stated you're not going to update IE's rendering engine 
without opt-ins (as opposed to the IE6->IE7 update), because too many 
sites would depend on your quirks -- so these are showstoppers, bugs 
that I'd call significant. Of course, it's not only the severity but 
also the plain amount of bugs that is important. If a fix breaks a few 
sites that can be adapted easily, that's feasible, especially if it 
helps other sites at the same time. That's how other vendors operate 
today. But the more such fixes hit a site negatively, the sooner it 
overwhelms the author, who will blame you then.


Received on Friday, 4 May 2007 18:40:28 UTC