Albert Lunde (
Wed, 28 Jun 95 12:17:07 EDT

At 10:15 AM 6/28/95, Terry Allen wrote:
>One wants to be able to mark up the contents of META (and thus it
>or some other element, to be invented, needs to allow content)
>because such undeniably common pieces of metadata as Title and
>Publisher may need to require such things as italics and superscripts.
>Even Subject could require superscripts.
>Albert's survey does suggest that META is currently in use; I would
>not call that "unsanctioned" as Stu did, though I agree with Stu that
>a small amount of use would not be an insuperable obstacle to changing
>the specification of META (*not* for 2.0 ...).
>However, it would be really bad design to create a new element to do
>what META doesn't quite do right. Maybe our process here doesn't allow
>us to do better, and much of HTML is already frozen for all time.

I'd suggest that <META> as currently defined and used is oriented towards
providing meta-data that might go in RFC822-like HTTP headers.

It's not totally off the wall to suggest a different tag/mechanism for
providing, say HTML-marked-up metadata or other arbitrary (text or binary?)

I'd also note that actual usage of <META> (with the arguable exception of
the "Refresh:" hack) is very much in the spirit of that suggested by the
syntax and examples in prior drafts. This is not a case like <BLINK> or
<CENTER> of something totally outside the standards process, this is a case
of software developers reading our drafts and taking them seriously.

The fact that we hit similar issues about style and applets (about the
limits of attributes as a place to stuff information), makes me wonder if
we shouldn't be advocating a more general way of hiding information (like
PIs or marked sections as suggest by Glenn) if we are talking about futures
and not HTML 2.0.

    Albert Lunde