RE: Versioning and html[5]

James Graham [mailto:jg307@cam.ac.uk]  wrote:
>Henri Sivonen wrote:
>> even
>> if we take as granted that IE needs opt-in versioning, there's no
>> guarantee that the opt-in flags for future IE versions will match the
>> publication of future HTML spec versions.
>
>I think this is a very important point. If IE7.5 and IE8 are both
>released in the HTML5 timeframe with IE7.5 introducing a HTML5 mode
>triggered off some assertation in the document that it is HTML5 (e.g.
>from the doctype), would this prevent IE8 from introducing any
>significant HTML/DOM/CSS features/bugfixes in HTML5 mode?

No.  But until the actual deployment of the new doctype reaches any critical mass (e.g. if it's <0.5% of the web, as "standards mode" was when we released IE6), we can change behavior without significant risk of breaking actual content on the web.

>It seems to me that development of language specifications and
>development of browsers are orthogonal processes that should not be tied
>together by versioning info.

They are orthogonal processes, yes.  But there is an interrelation.

>All future versions of the spec should be
>developed under the premise of "don't break the web" thereby making
>spec-version information in documents unnecessary*.

I agree with your premise (don't break the web) but not your conclusion (that spec-version information is therefore unnecessary).  Spec versioning will still tell you that your browser isn't new enough to handle, say, that 3d canvas element that gets added in html6.

>If UA vendors really
>believe that author opt-in is needed for all UA changes (and Microsoft
>seem to be the only vendor advocating this position), the only solution
>I can see is to force authors to specify their opt-in to the bugs and
>features of a particular UA  using e.g. <meta name="ua-version"
>value="IE7"> to specify IE7+ should work in IE7 mode when rendering the
>document. It's ugly as hell but it has the one redeeming feature of
>actually solving the stated problem which adding spec-version
>information to documents does not.

Even adding explicit versioning information will not take away IE's requirement to have a UA opt-in to "really standards" mode in IE.next.  However, having versioning in HTML will allow us to eradicate the need for authors to put this switch in every single document to opt in to good behavior, because we'll know that HTML6 content/apps won't expect to have the bugs we ship IE.next with.

-C

Received on Thursday, 12 April 2007 22:20:46 UTC