Re: META

Joe English (joe@trystero.art.com)
Wed, 5 Jul 95 13:31:37 EDT

Glenn Adams <glenn@stonehand.com> wrote:
> From: Joe English <joe@trystero.art.com>
> > 1. An unknown element encountered within HEAD shall be interpreted as a
> > implicit termination of HEAD and shall be interpreted according
> > to the content model:
>
> That heuristic will break on future extensions like <STYLE>,
> which can appear in the <HEAD> and has content.
>
> Oh? How would STYLE break this heuristic? STYLE only has #PCDATA content.
> It won't have any possible subelements which might be treated as unknown.

But STYLE isn't an HTML 2.0 element, so it would itself
be treated as unknown. (So would any other yet-to-be-invented
HEAD element with content.)

If the browser sees a (presumably legal HTML 2.1) document like:

<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML-NG Experimental//EN">
<HTML>
<HEAD>
<TITLE>Blah</TITLE>
<STYLE NOTATION=ARENA-STYLESHEET> <!-- unknown element in HEAD -->
...style specification...
</STYLE>
</HEAD>
<BODY>
...
</BODY>
</HTML>

the STYLE start-tag, since it's unknown by 2.0 browsers,
would by the above heuristic imply </HEAD><BODY>.
Then the current guidelines for dealing with unrecognized
elements would apply, and the "...style specification..."
would be rendered as plain text, which is undesired.

--joe@art.com