W3C home > Mailing lists > Public > www-html@w3.org > November 2002

Re: SGML SHORTTAGS Feature usage in HTML 4.x Recommendation.

From: William F Hammond <hammond@csc.albany.edu>
Date: 13 Nov 2002 08:36:58 -0500
To: Terje Bless <link@pobox.com>
Cc: W3C HTML Mailinglist <www-html@w3.org>
Message-ID: <i7ptt9a8w5.fsf@hilbert.math.albany.edu>

Terje Bless <link@pobox.com> writes:

> Bjoern Hoehrmann <derhoermi@gmx.net> wrote:
> 
> >* Terje Bless wrote:
> >>[The] HTML 4.01 Recommendation [...] prohibitions
> >                                      ^^^^^^^^^^^^
> >Recommendations at most.
> 
> Ok, if you want to nitpick... Since the text does not contain any "MUST" or
> "SHOULD" or "SHALL", what that text in a strict formal sense describes, is
> not quite an informational note. Do you really disagree at it's intent that
> these features be avoided? Especially in light of the change introduced
> with XML and XHTML 1.0?

I think this issue comes down to the question of whether HTML 4.01 is
an SGML application (and nothing more than that) or whether HTML 4.01
is its *own thing* with which there is a canonically associated SGML
application[1].

The latter is correct.  For example, one cannot make it a requirement
in an SGML application that there be a formal public identifier since
from the point of view of the strict SGML world an SGML document is
the full assemblage.

Therefore, I believe that the HTML 4.01 specification was erroneous in
the matter of the handling of SHORTTAG in its SGML declaration (but
I've not checked all the little detailed assertions made about how it
should be) if, at the time HTML 4.01 became a recommendation, all of
the necessary SGML infra-structure was properly in place.  (Were there
ISO actions still pending?)

It would then be unfortunate that documents which once were valid
might no longer be valid, but about such documents:

1.  They would not measure up in most past or present user agents.

2.  They can easily be translated to be compliant with the intent
    of the HTML 4.01 recommendation and with the formal SGML
    requirements of a corrected HTML 4.01 recommendation.

                                    -- Bill

[1] The same issue applies to the relationship between XML and SGML:
an XML application is its own thing.  There is an SGML application
that is associated with an XML application when an SGML declaration
for XML is used and an XML-compliant SGML document type definition is
referenced.
Received on Wednesday, 13 November 2002 08:37:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:53 GMT