David Morris (
Wed, 28 Jun 95 15:18:31 EDT

On Wed, 28 Jun 1995, Glenn Adams wrote:

> Date: Wed, 28 Jun 95 12:36:08 EDT
> From:
> And indeed, I *DO* want displayable metadata... I just don't want it
> displayed as a matter of course by browsers.
> If you want particular kinds of meta data to be displayable as content,
> then you should consider adding a new BODY element in the same vein as
> BANNER. IMO, META is the wrong vehicle for what you and Terry are suggesting.

This WG has long asserted that the content of <head> is not supposed
to be rendered as part of the document. There was a question earlier
in the last week as to whether this made it into the draft HTML 2.0
but if not we need to fix that as I believe there is long running
concensus of this group that head content doesn't get automatically
rendered as part of the document content.

UAs which don't conform will get deselected by customers when
<meta> or <style> data suddenly clutters up the display.

Use of <meta> as extended (or a new element) as Terry and Murray (and others)
have suggested should not cause users of conforming UAs any
difficulty. On the other hand, introducing a new element in the
<body> will cause difficulty since the default is to ignore
unknown tags and render the content.

Having the data renderable doesn't mean it is ever intended to be
rendered with the current document. It can clearly be data about
the document and hence meta data and yet need descriptive/ structural
markup. It will make much more sense to the user community if
that markup is HTML (or SGML).

Given the <head> content rendering rules, I've seen no argument that
would indicate any existing use of <meta> would be broken by
changing <meta> from empty to having content and an optional
ending tag. Have I missed something? If not, I see no reason
to invent a new element (which wouldn't solve the rendering
issue in any case) when it seems to me that <meta> is sufficient.

Dave Morris