W3C home > Mailing lists > Public > www-validator@w3.org > September 2005

Re: W3C Validator vs Schneegans

From: Henri Sivonen <hsivonen@iki.fi>
Date: Tue, 6 Sep 2005 20:53:37 +0300
Message-Id: <4c989197175343bf73bffda46d42c4f6@iki.fi>
Cc: www-validator@w3.org
To: Terje Bless <link@pobox.com>

On Sep 6, 2005, at 10:01, Terje Bless wrote:

> Christoph Schneegans <Christoph@Schneegans.de> wrote:
>
>> - It is unable to check XML well-formedness, but doesn't admit this.
>
> To the best of my knowledge, XML well-formedness checking is a strict 
> subset of
> validation. Thus to claim the validator can't do well-formedness 
> checking would
> be to misrepresent the case.

I think your reasoning is backwards. You seem to be saying that since 
the Validator is a validator, it must be checking for well-formedness. 
However, the reasoning should go the other way. The Validator does not 
check for all well-formedness constraints (proof by demonstration; see 
http://hsivonen.iki.fi/test/ill-formed-but-sgml-valid/ ). Therefore, 
the Validator is not an XML validator!

It would be nice if this was openly admitted in marketing instead of 
the "some limitations" euphemism in results.

>> Instead, it uses the euphemism "some limitations". Web browsers today 
>> are
>> able to find (almost all) well-formedness errors, the validator isn't!
>
> Please report these to Bugzilla (including test cases) if they aren't 
> documented
> allready. What we can't fix today will be very usefull as test cases 
> for the
> future.

I think it is futile to treat the individual problems as separate bugs. 
Rather, the Validator should use a real XML parser for XML.

> While some of them may be fixable with the current parser, the plan for
> addressing these shortcomings long term is to make use of a 
> specialized XML
> processor. This requires some fairly big changes in the code  which 
> is one
> reason why it's taking so long

XML has been around for over seven years and XHTML over 5 years during 
which time the Validator has had "some limitations".

> and is not without its own issues (determining
> when to use the XML processor and when to use the SGML parser, for 
> one).

Why not use the same heuristics that are now used for choosing the SGML 
declaration?

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Tuesday, 6 September 2005 17:53:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:20 GMT