W3C home > Mailing lists > Public > www-validator@w3.org > February 2001

Re: Validator should not be interpreting <SCRIPT> tag contents atall

From: Tim Bagot <tsb-w3-validator-0002@earth.li>
Date: Mon, 12 Feb 2001 09:51:50 +0000 (UTC)
To: <www-validator@w3.org>
Message-ID: <Pine.LNX.4.30.0102120921270.560-100000@windle.hack>
At 2001-02-08T09:05-0500, Bryce Nesbitt wrote:-

> Think of it this way.  Right now the validator "cries wolf", and
> therefore won't be used to validate real world JavaScript pages.
> What's better: intellectual purity or getting more people to validate
> more of their pages?

There is no point in more people validating their pages if the validator
no longer validates. The validator is intellectually neutral. Its purpose
is to attempt to parse a document as SGML, and signal whether or not it
conforms to the referenced DTD. It was never intended to check the other
constraints imposed or suggested by the various HTML specifications; it
should be considered as no more than just one of the tools an author can
use to find mistakes.

It can determine whether a web browser *ought* to render the document
correctly, but not whether it will, nor whether the document says what the
author intended. If you start adding options to emulate interpretation by
different browsers, why bother validating at all? Let us not condone poor
implementations of the standards by suggesting that their shortcomings are
reasonable variations.

> And there is a problem just as serious as getting the validator to
> stop complaining about non-problems.  Javascript is often used to
> BUILD html.  What's relevant is not the syntax rules of the bare
> document, but the syntax rules of the document after interpretation.

Both are relevant: the document must be valid both before and after
dynamic modifications. Checking the document afterwards is indeed a
problem. I suspect that proper testing can only be possible where dynamic
modification is not necessary. As soon as the document can be generated
only with the help of scripts, there will be too many variables for it to
be at all realistic for any validator to consider all possibilities.

Tim Bagot
Received on Monday, 12 February 2001 04:52:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 14:17:28 UTC