Re: HTML 2.0 reconstruction done

Roy T. Fielding (fielding@avron.ICS.UCI.EDU)
Fri, 31 Mar 95 05:53:40 EST

> I have to say I read this negligently, thinking it was the wording
> in the first version's 2.5:
>
>> When the above conflicts with the SGML standard, the SGML standard
>> may be ignored. Note, however, that not all HTML applications are
>> capable of ignoring the SGML standard.
>
> to which Gavin suggests the change
>
> When the above conflicts with the SGML standard, the SGML standard
> may be ignored. Note, however, that not all HTML applications are
> capable of ignoring the SGML standard, and that strict conformance
> may be required in the future.

I would accept "may be required for future levels of HTML".

> The old language was:
>
> 2.5 Understanding HTML and SGML
>
> HTML is an application of ISO Standard 8879:1986 -
> Standard Generalized Markup Language (SGML). SGML is a
> system for defining structured document types, and
> markup languages to represent instances of those
> document types. The SGML declaration for HTML is given
> in Section 5.1. It is implicit among HTML user agents.
>
> If the HTML specification and SGML standard conflict,
> the SGML standard is definitive.
>
> and that is the only approach I can support. HTML is defined as
> an application of SGML; we cannot ignore the SGML standard when we
> choose. It's that second para, saying that the SGML standard is
> definitive, that is still needed. The wording about how some HTML
> apps can't ignore HTML is kinda odd considering almost all of them do.

On the contrary -- we can and do ignore the SGML standard with a great
deal of regularity and for many good reasons. That is life! No user agent
is required to be an SGML application, and those applications are quite
capable of ignoring the SGML standard (sometimes in an unfortunate way).

> What's at issue here is how browsers are to do error recovery;
> let's not say we're defining an SGML app and then saying the SGML
> standard isn't normative for SGML apps.

We are not defining *just* an SGML app. We are defining a media type
that is both SGML-conformant and a reasonable proximity to what people
were calling "text/html" back in June of last year. That is why we
are in an IETF WG instead of an SGML Open group.

User agents can (and in some cases, should) bend the rules of SGML
in order to provide maximally robust interface to the user. Quite frankly,
this is an area that Internet people have had more experience with than
SGML people, and I think SGML folks should learn from it just like we
have learned the benefits of formally-structured documents.

On the Internet, shit happens on a regular basis -- a standard which
is not capable of coping with that (and in a consistant manner) is not
worthy of becoming an Internet standard.

....Roy T. Fielding Department of ICS, University of California, Irvine USA
<fielding@ics.uci.edu>
<URL:http://www.ics.uci.edu/dir/grad/Software/fielding>