Re: simple HTML list considered harmful

marca@ncsa.uiuc.edu (Marc Andreessen)
Date: Thu, 30 Sep 93 00:15:19 -0500
From: marca@ncsa.uiuc.edu (Marc Andreessen)
Message-id: <9309300515.AA20810@wintermute.ncsa.uiuc.edu>
To: Dave_Raggett <dsr@hplb.hpl.hp.com>
Cc: www-talk@nxoc01.cern.ch
Subject: Re: simple HTML list considered harmful
In-reply-to: <9309291523.AA23174@manuel.hpl.hp.com>
References: <9309291523.AA23174@manuel.hpl.hp.com>
X-Md4-Signature: 8f7db6815b422c4dc01109c380ec926d

Dave_Raggett writes:
> Marc writes,
> 
> > My objection is that people (real people -- users, authors, ...)
> > almost invariably think in terms of line breaks, not line beginnings.
> > Similarly paragraph breaks, not paragraph beginnings.  All (?) other
> > document processing systems (TeX, Microsoft Word, ...) hold to these
> > models.  Sure, you begin sections and you begin chapters and you begin
> > blockquoted regions, but paragraphs and lines are such a low-level and
> > intrinsic part of the concept and practice of text that it seems far
> > more natural to think in terms of breaks, as usual.
> 
> Defining <P> and <BR> as separators means you can't treat them as
> containers. 

Yup.

People want paragraph breaks and line breaks.  They may also want
containers, but I see no reason to blur the two together when breaks
are so intuitive and so frequently used in regular text handling
systems.

> This means you can't tag a paragraph/line with an identifier or
> rendering attributes such as alignment.

Why not have containers for doing those things, as now, and keep them
away from the break concept and tags?

> Quite a few people have asked for <P> as a container and most word
> processing applications use a container model. 

Internally!  They don't ask you to *think* in terms of a container
when what you really want is a -- yup -- line break.

> There seems to be a growing feeling that many authors will create
> and maintain documents in a richer format and convert to HTML/HTML+
> for delivery to the Web. 

Experience is showing that the two methods for generating HTML are
this, and writing by hand.  Both happen, and both will continue to
happen for a long time.  It's (in a sense) irrelevant what the
underlying format looks like for the former case, but for writing by
hand, I think we should try to make sure it's reasonably intuitive --
HTML isn't too bad (but not real good either) currently, and moving
away from the concept of breaks where breaks are clearly what users
want in many cases would be a step backwards.

Cheers,
Marc