Re: Parsing methods

Paul Prescod (papresco@calum.csclub.uwaterloo.ca)
Thu, 11 Jul 1996 11:12:43 -0400


Message-Id: <1.5.4.32.19960711151243.0070ee88@csclub.uwaterloo.ca>
Date: Thu, 11 Jul 1996 11:12:43 -0400
To: Jeremy Bailin <j.bailin@utoronto.ca>, www-html@w3.org
From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
Subject: Re: Parsing methods

At 12:31 AM 7/11/96 -0400, Jeremy Bailin wrote:
>The problem is that humans are lazy. If you're hand-coding something, and you 
>know you can get away with certain constructs, and that every browser will 
>understand those constructs, you're likely to use them. Ditto if you're
writing 
>something that outputs HTML. In any case, if the browsers' behaviour is
erratic 
>in the face of invalid HTML, it has no implied validity. If the browsers' 
>behaviour is consistent and uniform across browsers, there is an implied 
>validity to what really should be an invalid construct.

It seems we go over this every few weeks. Can we agree that the best
behaviour is for the browser to provide a mechanism for _reporting_ invalid
documents but "do its best" to render it?

 Paul Prescod