RE: FW: I-D ACTION:draft-connolly-text-html-00.txt

On Tue, 5 Oct 1999, Larry Masinter wrote:

> > > If the HTML 4.01 document is inadequate, then we can and should fix it.
> > 
> > The one really important fixing it needs is to drop all normative
> > references to ISO8879.  [...] That, IMHO, would go a long way towards
> > making 4.01 an honest spec.
> 
> I suppose I should have qualified 'inadequate', since of course, there's
> no belief that the HTML 4.01 specification could be adequate for all
> purposes. In particular, it doesn't seem to purport to, or want to,
> document current practice. 

That's part of the fundamental problem, really.  The spec has no grounding
in practice.  Instead, it's cast in a formalism the impedance mismatch
between which and current practice couldn't be any more total. 

> I can note in the revision of the 'text/html' document that current
> practice on the Internet includes much non-compliant behavior, and
> that implementors must be prepared to be "bug-compatible" with popular
> browsers in order to work with many HTML documents on the net.

This is the other part.  Such alleged non-compliant behavior is in
relation to a formalism which few if any implementations have made even a
token effort to abide by.  We can sneer at Tag Soup, but that can't and
doesn't deny it a place under the sun, or the possibility - never mind the
desirability! - of a rationalizable basis of its own. 

In the thread from the HTML-WG that I cited, I had written (quoting Dan
Connolly):

:> I suggest we approach this education problem -- which is what is is --
:> from the top down: educate the implementors of browsers and authoring
:> tools, and the authors of HTML "how-to" documents. Let the consumer
:> community learn from them. That's why I stared the HTML 2.0
:> specification effort in the first place.
:
: I'm sorry to disagree, but what you have here is actually a
: *re*-education problem. The explosive growth of the Web mainly comprises
: people who come from a different software universe/paradigm: closed
: proprietary programs, misleading manuals, a thriving genre of "tips,
: tricks and traps" books, and so on. The common theme is that the
: *program* is the *final* arbiter of what's OK. People learn from and
: adapt to the implementation. There's no such thing as "the program is
: doing this wrong."

Four years later, I stand by that statement, and I believe the IETF's
desire to wash its hands off HTML is conclusive. 

So, I submit that "non-complaint behavior" and the advisability of being 
"bug-compatible" are unfair and misleading characterizations of the
reality of text/html.

  1. The markup practices deployed in the corpus of documents "out there"
     reflects the fact that people have learnt from and adapted to certain
     popular implementations.
  2. These implementations embody a somewhat "streamed" processing model
     in which, as a rule, tags in isolation are associated directly with
     obseravble rendering consequences.  ("See a tag, do something; no
     tag, no action".) 
  3. While SGML has been (a) the original inspiration for the HTML tagset,
     and (b) useful in providing a formal vocabulary to express certain
     desirable structural and semantic features, it has neither any
     bearing nor any relation to both the theory and practice embodied
     in the popular implementations.
  4. Interoperability with popular implementations is vox populi.

I think the revised statement on text/html should reflect rather than
suppress the import of these propositions. 
 
> I don't think this reduces the value of specifying what 'text/html'
> *should* be, although I agree it makes implementation hard.

A spec is allowed to elaborate on recommended practice.  I don't think
implementing tag-soup processors is hard.  The argument against tagsoup
proceeds on a different level: that it's a limited and dead end paradigm.
But this does not obviate an obligation to document practice faithfully,
if such a document needs to exist *at all*. 

> If you know of good published material (with stable references)
> for explaining the "real HTML", could you send me the bibliographic
> information?

I don't know about "good", but a search engine query for something like
"HTML Tutorial" will surely return a wealth of material, on a par with
this recent posting to alt.html, quite sanguine in its take on practice:

 <URL:http://www.deja.com/=dnc/getdoc.xp?AN=527653317>   



Arjun

Received on Wednesday, 6 October 1999 02:22:46 UTC