Re: What are the problems with IDML? (fwd)

Daniel W. Connolly (connolly@w3.org)
Mon, 26 Aug 1996 01:47:44 -0400


Message-Id: <199608260547.BAA10722@anansi.w3.org>
To: donohoe@emerge.com
cc: Murray Altheim <murray@spyglass.com>, www-html@w3.org
Subject: Re: What are the problems with IDML? (fwd) 
In-reply-to: Your message of "Fri, 23 Aug 1996 09:32:54 +0800."
             <321CFC36.39EB@emerge.com> 
Date: Mon, 26 Aug 1996 01:47:44 -0400
From: "Daniel W. Connolly" <connolly@w3.org>

In message <321CFC36.39EB@emerge.com>, Doug Donohoe writes:
>
>This may be a stupid question but....

I'm afraid I have to agree with you there ;-) Please read on...

>If I put META in the body of my HTML document, neither Netscape
>nor IE choke (they seem quite complacent about it, as a matter of
>fact).  So, what exactly _does_ break when I put
>META inside the BODY?  If 99% of HTML is viewed through these
>two browsers, who cares if I put META in the "wrong" place?
>I know SGML or HTML parsers will complain, but what does
>it break in the real world?  E.g., what problems does it cause?

It doesn't cause any immediately observable problems. At least
I don't think so. Being a commercial developer, you'd better
test this conjecture against all the relavent products to be sure!

But disregarding the specs -- and this is assuming you would disregard
the specs, rather than negotiating in the appropriate forums to get
them changed. I know _you_ wouldn't do that :-), but for all the folks
out there considering it: it is destructive to the fabric of the web
and the net in the following way:

It breaks down trust. It breaks down that notion that the net and the
web are built on public specs -- that it's a level playing field out
there, cuz anybody can go and read the specs and implement them.

If the documents out there don't conform to the spec, then folks
can no longer just code to the specs and expect it to work. They
have to reverse engineer the behaviour of the market leader(s)
de-jour.

So please, everybody, stop asking "what would it break?" and start
asking "what is the design rationale behind the current spec, and is
there new evidence that suggests the spec should be changed?"

Once you've got an argument that the spec should change, then we'll
all start to look real hard at compatibility and deployment issues.

Some might say that this battle has already been lost -- that
coding to the specs is a losing proposition anyway. I maintain
that it is cost-effective in the long run, and I will continue
to choose to see the world that way for as long as I live, or
until I become a cynic like everybody else :-{

Dan

p.s. All the folks with the knee-jerk reaction about how W3C specs
aren't public, just save it, OK? I'm not in the mood to apologize for
the fact that the relavent players aren't willing to duke it out in
public forums -- public forums where folks aren't disciplined enough
to research their questions and cite sources in the answers. In
fact, I'm in a pretty foul mood all around >-{