Re: Differences between validator output and nsgmls output

On Wed, 4 Aug 2004, Jirka Kosek wrote:

> But in this light wouldn't it be better to
> disallow SHORTTAG option in SGML declaration for HTML in the validation
> service?

That's what many people think, or they at least think that it is good to
have two "parsing modes". But HTML, as defined in HTML 2.0 through
HTML 4.01, allows SHORTTAG.

> HTML 4.01 specification doesn't mention possibility of using
> SGML markup minimalization rules AFAIK.

It's not very explicit, for obvious reasons (its authors knew that
browsers hadn't bothered implementing SGML rules), but since it
definitely defines HTML as an SGML application with a declaration that
allows SHORTTAG, the normative side of the matter is established even
without any mention in understandable prose.

Besides, it _does_ mention minimization, though somewhat oddly, in
(non-normative) appendix B. At
it says:
"B.3.3 SGML features with limited support
SGML systems conforming to [ISO8879] are expected to recognize a number
of features that aren't widely supported by HTML user agents. We recommend
that authors avoid using all of these features."
It does not explicitly say what those features are, but presumably
subclauses 3.5 through 3.7 are what it refers to (and should therefore
appear as subclauses 3.3.1 etc.). The feature under discussion is
in subclause 3.7.

The situation is pretty odd. It is a normative part of the HTML 4.01
specification that the feature is allowed, so browsers are required
to support it - but hardly anyone thought the vendors would take this
seriously. Authors are, non-normatively, advised not to use the feature.
Well, most authors never heard of it, so they only encounter it in
validators' error messages, in cryptic ways.

Jukka "Yucca" Korpela,

Received on Wednesday, 4 August 2004 12:43:25 UTC