Re: [Fwd:]

On Tue, 3 Aug 2004, Nick Kew wrote:

> If the author selects a parse mode better-suited to HTML they will get a
> less-confusing error message.

A parse mode that allows no tag minimization might be "better suited" to
most people's (including people who created Web browsers) understanding
of HTML, but it's still contrary to HTML specifications (from HTML 2.0 to
HTML 4.01), which specifically allow minimization. But we've gone through
this, haven't we?

> >	 Few authors understand that e.g.
> > <a name="foo"/>/
> > is perfectly valid HTML4.
> Huh?  We all know that's perfectly valid (though of course it has to be
> closed).

It _is_ closed. Naturally it needs to be included into a suitable context,
an HTML document, e.g. the four-liner

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
<a name="foo"/>/

Both the <a> tag and the <a> element _are_ closed in the snippet I gave.
It contains more than a complete <a> element, not less. By HTML4 rules,
that is.

Regarding the idea that XHTML 1.0 restricted as per appendix C and
served as text/html is the practical approach, it gains nothing as
compared with HTML 4.01 in treatment by browsers. In validation,
you avoid some problems and create some new. For example, if you
accidentally write <textarea name="foo" rows="5" cols="*">,
a validator will report a syntax error by HTML 4.01 rules, but not by
XHTML rules. (In XML, and hence in XHTML, a restriction that an
attribute value must be an unsigned integer cannot be expressed
in a DTD, hence cannot be checked by a validator.)

Jukka "Yucca" Korpela,

Received on Tuesday, 3 August 2004 01:31:24 UTC