Re: Can anyone spot the bug?

Paul Prescod (papresco@calum.csclub.uwaterloo.ca)
Thu, 11 Apr 1996 13:19:00 -0400


Date: Thu, 11 Apr 1996 13:19:00 -0400
Message-Id: <199604111719.NAA04336@undergrad.math.uwaterloo.ca>
To: bbreak@mit.edu (Ben Breakstone)
From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
Subject: Re: Can anyone spot the bug?
Cc: ao950@freenet.carleton.ca (Paul Derbyshire), www-html@w3.org

At 12:45 PM 4/11/96 -0400, Ben Breakstone wrote:
>Um... this mailing list is about the design of HTML.  This may sound
>didactic, but I don't know that questions about the operation of Netscape's
>frames and scripting language belong here, as they aren't actually part of
>HTML.  There's been a lot of traffic on the list lately about those topics.
>Perhaps discussion of the inclusion of these features in HTML....

I agree 100%. I sent private mail along the same lines. 

>I can't speak for any other readers, but I'm not interested in Netscape's
>scripting features in the least- I've found them unstable, inconsistent,
>and poorly-designed.  

I agree here, too. 

> But that's a topic for another list.  ;-)

But not here. Scripting features in HTML _is_ a topic for this list, and we
can learn how _not_ to do it by examining JavaScript. In particular, my
complaints are:

* Hidden from SGML tools -- very fragile. Some tools might delete comments
to save bandwidth.
* Outside of the SGML document structure -- loses all SGML structuring benefits.
* Can change SGML document's structure -- not only renders existing
validators useless, but renders new validators impossible to right (the
halting problem)
* Not properly integrated with other parts of the system. Sometimes going
"back" from a link will take you back to a JavaScript function! Weird.

  Paul Prescod