On Tue, Jul 14, 2009 at 9:02 AM, Simon Pieters<sim...@opera.com> wrote: > On Tue, 14 Jul 2009 14:51:42 +0200, Tab Atkins Jr. <jackalm...@gmail.com> > wrote: > >> How do y'all currently handle <noscript> content in a context that >> allows scripts? What if there was a <video> or <object> in the >> <noscript>? > > <noscript> is parsed as a CDATA element in the HTML parser when scripting is > enabled. Any tags would be interpreted as text. > > This is not the case for <object> or <video> because it's ugly and doesn't > work in XHTML. You'll notice that <noscript> is not allowed in XHTML5.
Hrm, that's no good then. *Something* has to be done about it, though, because just display:none'ing the fallback content doesn't allow legacy video clients to be tried, which the spec explicitly mentions (and just plain makes sense anyway - without it, <video> adoption will be substantially slower). We need *some* way to indicate that particular elements aren't to be activated at all in certain circumstances. On a side note, does the old <object><embed></> trick work only because the browsers that support that use of <object> don't support that use of <embed>, and vice versa? So even though the <embed> is parsed in an <object>-supporting browser, it doesn't activate? If so, then it seems like our problem is *too much* interop now. ^_^ ~TJ