Re: Abstraction in HTML

Brian Behlendorf (
Mon, 12 Dec 1994 11:48:41 -0800 (PST)

Date: Mon, 12 Dec 1994 11:48:41 -0800 (PST)
From: Brian Behlendorf <>
To: "John C. Mallery" <>
Cc: Multiple recipients of list <>
Subject: Re: Abstraction in HTML
In-Reply-To: <>
Message-Id: <>

On Mon, 12 Dec 1994, John C. Mallery wrote, quoting Michael Johnson:
>     Not good idea. For one thing, other GML implementations also have UL and OL
>     and DL, so for consistency from one GML to another, lists should be left this
>     way.
>     I also do not think this would make HTML any easier to write or understand,
>     quite the contrary, I think it would make things less clear. 
> People should not be writing html; programs should.


Which ignores the whole reason why HTML became as widespread as it has.  It's
precisely *because* HTML was easy to write *by*hand* that techies and
non-techies alike could create the thousands (millions?) of web documents out
there.  Now I haven't seen the MS Word "Internet Mode"  extensions, but I'd
be very surprised if it allowed people to write HTML without ever seeing an
HTML tag or understand HTML semantics.  If we are going to encourage people
to mark up their documents semantically, we have to get away from the notion
that knowing HTML semantics is a bad thing.  Why would someone put a header
in <H1> when they could simply put it in a larger font?  Or use <address>
when they could use <em> or <i> or put "italic" in their stylesheet?  
"Save as HTML" in a WYSIWYG authoring environment will almost always 
result in bad HTML unless semantics are a core part of the creation 

HTML 3.0 is just about right on that - it's kept relatively simple.
There are some parts, like tables and embedded imagemap functionality in 
FIG and possibly maths that I would expect people to use a tool to 
create, but tags like <font> couldn't be created in anything but a 
WYSIWYG authoring environment.

When considering changes to HTML, let's focus on why it was this language
that became the preferred format, and not RTF, LaTex, or PostScript, and
let's make sure we keep HTML from going down those paths.  HTML can, at best,
represent a higher "idea quantum" level than plain unmarked text; the highest
level being the ideas in the authors head, the lowest level a language like
PostScript.  This is why I strongly favor using stylesheets - they allow the
HTML document to remain simple, semantic, and flexible, while giving those
who care a lot about presentation to express their preferences. 



Your slick hype/tripe/wipedisk/zipped/zippy/whine/online/sign.on.the.ish/oil
pill/roadkill/grease.slick/neat.trick is great for what it is. -- Wired Fan #3