W3C home > Mailing lists > Public > www-html-editor@w3.org > July to September 1997

The Problem With <NOSCRIPT>

From: Rev. Bob <rev-bob@full-moon.com>
Date: Mon, 14 Jul 1997 23:19:30 -0400
Message-ID: <33CAEC42.109DE0A8@full-moon.com>
To: hwg-newtech@hwg.org, hwg-html@hwg.org, www-html-editor@w3.org
Has anyone else noticed the fundamental flaw in the <NOSCRIPT> tag as
currently implemented?  Consider the following page (sans DOCTYPE):

<title>Document Title</title>
<script language="Foo">Boogedy!<br></script>
<noscript>What language is "Foo?"<br></noscript>
End of text.

In theory, this is easy to follow; since there is no "Foo" script
language, the <script> content will be ignored and the <noscript>
contents will be shown.  Thus, the display should be:


What language is "Foo?"
End of text.


However, that ain't how it works.  Try it in MSIE 3.x or Communicator
and you'll see absolutely NO output...either way.  Why?  Because while
they don't recognize "Foo" script (thus skipping the <SCRIPT> tag), they
parse <NOSCRIPT> contents *only if scripting is disabled*...whether they
understood the <SCRIPT> tag that immediately preceded it or not.

The HTML 4.0 draft specification only makes matters worse.  It
depreciates the use of the LANGUAGE attribute for <SCRIPT> in favor of
the TYPE attribute.  Y'know, JavaScript 1.0, 1.1, and 1.2 ALL share the
"text/jss" MIME type!  THIS DOES NOT HELP.

Here's what I propose.  Before the LANGUAGE attribute can be
depreciated, it needs a replacement that will adequately substitute for
the information that it provides.  To that end, I propose allowing the
<NOSCRIPT> tag to take the TYPE attribute, as well as developing a new
VER (version) attribute that would apply to both <SCRIPT> and
<NOSCRIPT>.  With those changes, the above page would look like this:

<title>Document Title</title>
<script   type="text/foo" ver="1.0">Boogedy!<br></script>
<noscript type="text/foo" ver="1.0">
What language is "Foo?"<br>
End of text.

In this model, the <NOSCRIPT> tag is activated if the browser does not
understand the script language specified in the type/ver attrbute pair -
in this case, FooScript 1.0.  The code will still degrade well on
non-scripted browsers (identically to the current model), and would
operate better on future browsers.  Unfortunately, nothing short of
manufacturer patches will fix existing browsers.


 Rev. Robert L. Hood  | http://rev-bob.home.ml.org/
   Rage InfoCentral   | http://rageccg.home.ml.org/
   CCG InfoCentral    | http://rev-bob.home.ml.org/info_cen.html
 Chattanooga, TN LARP | http://paradigms.home.ml.org/
  Get Off The Cross!  | http://gotc.base.org/
Received on Monday, 14 July 1997 23:16:50 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:08:19 UTC