Re: [xml-dev] XMLisms and HTML parsing and modes (was: Re: XML5)

[Liam R E Quin <liam@w3.org>]

On Thu, 2010-12-16 at 15:38 -0800, Henri Sivonen wrote:
> On Dec 16, 2010, at 14:13, Liam R E Quin wrote:
> 
> > This <syntax /> came from SGML and was already legal in HTML from a
> > formal standards perspective.
> 
> My understanding is that Annex K was added to SGML in late 1997 for
> the specific purpose of making XML 1.0 a subset of SGML. This was
> after HTML had been deployed.

Yes. But the net syntax was already in SGML:1986.

> Also, I believe no IETF or W3C HTML spec that pretends HTML to be an
> application of SGML actually enables the NET-enabled start tags.
> 
> Am I mistaken?

I'm not sure about "pretends to be HTML". SGML allows any *document
instance* to enable them.

However,
http://www.rfc-editor.org/rfc/rfc1866.txt
has
    SHORTTAG YES
which results in enabling the NET feature.

So it was in HTML at least as far back as 1995.

It was probably done so that attribute values didn't need to
be quoted as long as they parsed as SGML tokens, but had the
consequence of enabling a whole bunch of other <i/features/ :)

The SGML spec, and the SGML declaration, are both extraordinarily
cryptic in this area.  The SGML Handbook does help, but annotated
mud is still mud.

We're a bit far afield really, except that it's not correct to
blame XML for introducing a syntax that was arguably already a
part of HTML (and accepted by at least some tools).

Liam

-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/
Ankh: irc.sorcery.net irc.gnome.org www.advogato.org