Re: end tags...

Murray Altheim (
Wed, 25 Sep 1996 12:05:05 -0500

Message-Id: <v02140b0bae6f02f5f025@[]>
Date: Wed, 25 Sep 1996 12:05:05 -0500
To: Lee Daniel Crocker <>
From: (Murray Altheim)
Subject: Re: end tags...

Lee Daniel Crocker <> writes:
>> >readable/writable language?  Most poeple who hand write do it
>> >litterally to make sure that when the rules NEED to be broken they can
>> They NEVER *need* to be broken IMHO.
>Of course they do.  If you want clean information, you have to
>take out things like physical markup and put them in style sheets--
>but then your code won't validate, because the "CLASS" attribute
>is not part of any existing DTD.  Of course we all know it will
>be in the next one, but for now, you can write cleaner code by
>violating the standard than by following it, and should, IMNSHO.

Not true. You can use CLASS and ID in the i18n DTD, or Internet Explorer's.
And there's nothing wrong with creating your own DTD. The question is UA
support. Browsers should ignore elements and attributes they don't
understand, so if your DTD operates transparently and your DOCTYPE points
to your DTD, you're legal. But I agree in principle to your main point --
it's unfortunate that the development in W3C has removed any priority for
activity in the IETF, which is The standards body in this neck of the

>I also sympathize with complaints that HTML is [:]
>10 times more complex than necessary,

I think it's complex in ways it needn't be. If stylesheets had been
available a year earlier, we wouldn't have seen rampant use of presentation
elements, which has turned into the de facto method of implementing RTF in
HTML. Really, really ugly. And HTML is missing potential structural markup
elements, which is only now being hacked in using things like DIV and SPAN,
rather than actual document features like TOC, HEADER, FOOTER, etc.

Sunil Mishra <> writes:
# So, HTML should ideally be written independent of the browser (user agent,
# actually) which is to read it. [...] It's a pity that a lot
# of people choose to mangle structural components to try and obtain a look,
# that is what CSS is for. And you can blame Netscape and MSIE for dragging
# their heels on this. (Well, maybe not MSIE as much.)

Looking at their latest authoring tools, it's still <FONT FACE= SIZE=
STYLE= BLINK= BARF= BARK= ETC=> all the way. I was a little surprised at MS
Internet Assistant, as I thought all that stuff would have taken advantage
of MSIE's stylesheet feature. Next version, maybe.

>hopelessly ambiguous, unparseable,

Disagree here -- not ambiguous at all, and certainly parseable. The
specification is pretty tight (and you could make it tighter), and
validators parse it just fine. I don't catch your drift here I suppose.

>based on a spec that isn't even legally publishable on the Web,

There are many implementations of technologies whose specifications are not
available for Web publication (ie., essentially in the public domain). The
SGML specification is not free nor public domain, but it is publicly

>and not as expandable as it could be.

It takes time to move forward. It's certainly the most expandable and
platform-independent document format in wide use. Wait until we see Java,
JavaScript, VBScript, CORBA and ActiveX all supported on the Web. That will
of course be an enormous mixed blessing, going back to your point about
being overly complex.

>That doesn't mean I advocate
>ignoring it, but let's not sweep all the complaints under the
>rug and call the complainers uneducated; let's fix it.

Hey, we're all trying.


     Murray Altheim, Program Manager
     Spyglass, Inc., Cambridge, Massachusetts
     email: <>
     http:  <>
            "Give a monkey the tools and he'll eventually build a typewriter."