Re: HTML/XML Task Force Minutes 11 Jan 2011

John Cowan wrote:
> >    HS: So, the pattern is formalized in HTML5. An alternative is
> >    using
> >    <STYLE>.
> 
> Does that really work?

Yes:
http://livedom.validator.nu/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%20type%3Dnot%2Fknown%3E%0A%3Cfoo%20xmlns%3D%22http%3A%2F%2Fhsivonen.iki.fi%2FFooML%22%3E%0A%3Cbar%2F%3E%0A%3C%2Ffoo%3E%0A%3C%2Fstyle%3E

The tokenization of the content of <style> is not the same as the tokenization of the content of <script>, though. <style> is less magic.
http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#rawtext-less-than-sign-state
vs.
http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#script-data-double-escaped-less-than-sign-state

> IIUC, descendant elements of STYLE are parsed
> and
> then the value (in the XSLT sense) is taken. So it wouldn't work for
> embedding unescaped XML, because the XML would wind up being parsed as
> HTML.

That's not how it works.

> >    <hsivonen> existing browsers wouldn't honor NORUN
> 
> True, but that matters only for a few media-types that might be run.
> In particular, it's safe to say that text/plain would never be run (in
> which case the "what to do with it" could be recorded in one of the
> data-*
> attributes or other extensibility points).

I think Noah's later observation that since NORUN would do nothing in browsers that don't recognize the value of the type attribute authors would fail to use NORUN at random is a much stronger reason against NORUN.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Thursday, 13 January 2011 08:24:20 UTC