Re: Conformance ratings (was: Extra! Microsoft beats Netscape

Harold A. Driscoll (harold@driscoll.chi.il.us)
Tue, 13 Feb 1996 14:33:40 +0000


Message-Id: <2.2.16.19960213143340.233f0eec@pop.interaccess.com>
Date: Tue, 13 Feb 1996 14:33:40 +0000
To: MegaZone <megazone@livingston.com>
From: "Harold A. Driscoll" <harold@driscoll.chi.il.us>
Subject: Re: Conformance ratings (was: Extra! Microsoft beats Netscape
Cc: www-html@w3.org, jonsm@aol.com

At 18:26 12/2/96 -0800, MegaZone wrote:
>>> Having error counts displayed in the search engines should give the authors
>>> incentive to clean up their act.
>
>I for one wouldn't be swayed by this at all.
>
>First of all, I've yet to see a validation system stay up to the minute
>with new extensions.  I don't think it is at all fair to penalize pages that
>use extensions if the underlaying code is solid.

Two of the most popular, WebTechs and Kinder-Gentler Validation will
certainly do so. They each use an SGML browser and a DTD. [1]

The problem is that the "extensions" to HTML are rather casually described
"in vague English" but are lacking in serious documentation such as a DTD.

As Abigail pointed out, for best results one needs both SGML validation and
some semantic checks of the "why would anything do such a thing?" nature.
WebLint is an example of such a tool, one I've found quite useful.

>And I don't expect anyone to start putting DOCTYPE notes in their pages.
>I'm not going to, it's a pain in the ass to keep that stuff straight.
>And since it shouldn't be necessary for browsers, why should I?  If a
>browser doesn't recognize a tag or attribute, it ignores it.  Why should it
>care about the DOCTYPE?>

Check out the discussion of the <!DOCTYPE...> statement in the RFC 1866 HTML
2.0 Specifications. You'll see that it is part of every properly compliant
HTML document. (True, you can often get your pages to work without one. But,
then, I've a friend who never changed the oil in his car. Worked fine for
him, for about 90,000 miles. True, he had to scout the neighborhood, and
drive from the down-hill direction when he traded it in, and very carefully.
Things worked out very differently with his next car, but that is quite a
story in itself.)

>The web is an open playing field.  If I decide to use <FONT COLOR="#rrggbb">
>on my pages, that's my decision.  I know very well that only NS 2.0 will
>use it.

Some folks will quite possibly be impressed, some may be bedazzled, and
others will likely be annoyed when things don't work with their browser.

A language is a tool, whether it is used to communicate with people or with
machines. One can be creative with spelling and grammar--let the reader
figure out what the author is having difficulty (or being difficult about)
saying.

A few accomplished writers have become very successful using such devices.
Most, however, present themselves as inept and unprofessional jerks. The
same is the case, whether it is spoken language (eg. English) or HTML.

Like you say, it is an "open playing field." So play away.

/Harold

[1] I've links to WebTechs and Kinder-Gentler Validation at
http://www.ccs.org/validate/ . As an aside, this is also an example of a
page which uses about a half-dozen different forms, a point which relates to
another thread.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Harold A. Driscoll                       mailto:harold@driscoll.chi.il.us
#include <std/disclaimer>      http://homepage.interaccess.com/~driscoll/