Re: HTML3 <OL inherit> gone for good?

Brian Behlendorf (brian@wired.com)
Mon, 13 Mar 1995 08:40:40 -0800 (PST)


Date: Mon, 13 Mar 1995 08:40:40 -0800 (PST)
From: Brian Behlendorf <brian@wired.com>
To: lilley <lilley@afs.mcc.ac.uk>
Cc: www-html@www10.w3.org
Subject: Re: HTML3 <OL inherit> gone for good?
In-Reply-To: <18921.9503131319@afs.mcc.ac.uk>
Message-Id: <Pine.BSD.3.91.950313080837.23952t-100000@get.wired.com>

On Mon, 13 Mar 1995, lilley wrote:
> Brian wrote:
> > On Fri, 10 Mar 1995, Lou Montulli wrote:
> > > Stylesheets should
> > > in no way be relied upon as the only method of rendering hints or
> > > assumed to exist before they actually do.  
> > 
> > In some ways you're right, this is a chicken-and-egg situation.  But 
> > there's sufficient evidence to suggest that stylesheets should at least 
> > be investigated by any institution doing serious R&D in web technology 
> > development.
> 
> Ok, Brian, so you are agreeing with Lou that stylesheets should, in time, be 
> implemented but that they should not be the only way to provide rendering hints?

I was implying that a company whose vision of furthering HTML development 
as nothing more than adding new tags and attributes left and right should 
reconsider.  If they're serious about conducting development as an 
long term experiment and not as a way to make the "new features for 
version x.x" list longer, then they should consider all solutions.  

> > No, the standards process must *not* be driven by market forces or the
> > momentum of first-to-market, and it should not be John Q. Public who decides
> > issues of fundamental web architectures.  
> 
> So, you are saying that document authors should not be consulted and no
> notice should be taken of what people want from the Web; instead, what is good
> for them should be given as a fait acompli and they can take it or leave it?

No, you're putting words into my mouth.  Of course the needs of authors 
should be paramount in the specification, and along with that the 
knowlege gained from having tried many different routes and finding the 
best solution.  What I'm worried about are entropic market forces, like 
"well, everyone is using IMG align=left, so we better put it in the 
spec".  There is not necessarily a correlation between widespread use of 
a thing, and the intelligence of a thing.

> I find it hard to believe that you are putting forward this line of reasoning.
> What will you do if document authors decide to 'leave it'?

My concerns entirely.  Believe me, I'm usually a populist - I firmly believe
that legitimacy lies in consensus.  However, I also know that sometimes a
majority of people can be wrong, and that certain principles must guide those
who lead in those situations - read up on the history behind the drafting 
of the U.S. Constitution and you'll see that these concepts are a lot 
more benign than they might sound.  If popularity were all that mattered, 
we would be declaring MS Word files the default web document standard ;)

> This seems a pretty radical departure from the HTML 2.0 philosophy, which
> rounded up all the things that were being both implemented and used by
> the HTML authoring community and produced a DTD that represented current 
> practice.

Well, those drafting HTML 2.0 didn't have much choice - in 1993 (2?) marca
added <IMG> and the rest was history.  Of course now we're paying because 
<IMG> should have been a container, as now <FIG> is in HTML 3.0 - which 
is a very good example of something we should aim to *not* have happen 
again.

> Brian, I hear you saying that stylesheets are great and have a lot of 
> potential. I agree. I suspect Lou does too.
> 
> But I don't see the problem with the existing stylistic hints that HTML 3.0
> provides (?used to provide?).
> 
> For example, people at our site have been putting in centered headings with 
> the Netscape specific <center> tag. I have been explaining to them, with 
> some success, that if they do it the HTML 3.0 way like this:
> 
>   <h1 align="center">
   
The current HTML 3.0 draft still allows this, and I'd argue to keep this 
(look at <URL:http://www.hpl.hp.co.uk/people/dsr/html/paras.html>).  This 
is all about how future development should move - should we continue to 
add align= attributes to every body tag one by one, and then color, and 
then size, and then font style.... or should we generalize and 
*ironically enough save ourselves some effort* and go for style sheets?
We need to consider the mess we would be in if people misinterpreted 
experiments as features (as many are now) and we went crazy with 
adding HTML attributes.  

	Brian 

PS - relevant quote du jour: "Incrementalism is innovation's worst enemy" 
- Nick Negroponte

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@hotwired.com  brian@hyperreal.com  http://www.hotwired.com/Staff/brian/