RE: background-color not validating when it should

I was not trying to disagree with David's logic. I believe Philip explained
the point better than I did.

I agree 100% with electing to only receive such warnings through the
advanced interface. I think that the default behavior should be only to
validate. If we want warnings, electing to see them outside of the default
check would be a nice alternative.

We include validation buttons on many of our client sites so they can rest
assured we are adhering to the standards. Unfortunately, though not
critical, warnings make them think something is wrong even though it is not.

Thanks,
Robert Mullaney
Easy Business Services
Phone: 239-242-6691 x 22
 
-----Original Message-----
From: Disgusted of Tunbridge Wells
[mailto:Disgusted@Royal-Tunbridge-Wells.Org] 
Sent: Sunday, November 19, 2006 4:57 PM
To: Robert Mullaney
Cc: 'David Dorward'; www-validator-css@w3.org
Subject: Re: background-color not validating when it should



Robert Mullaney wrote:

> A 2-sided argument is pointless. I would not have posted this if I knew
only
> 1 person would get involved in the discussion.
> 
> Doesn't anybody else have an opinion on this? Obviously we know my point
of
> view and David's. I find it impossible to believe nobody else has anything
> to say.

I have an opinion.  On this point at least, David's logic
is flawless.  Either the validator is entitled to issue
warnings or it is not (/tertium non datur/ applies here
without qualification).  Since, by definition, anything
that is incorrect as per the CSS spec. must generate an
error, warnings -- if they are to be issued -- must be
reserved for those cases wherein the spec. is met but
caution is nonetheless needed.  This is clearly one of
those cases.

The question is therefore "Should the validator issue
warnings", and I believe that the answer is "no".  We
have been told time and time again that the validator
(usually the HTML validator, but the argument applies
equally to the CSS validator) is not a LINT-like tool.
It should therefore stop behaving like one, at least
in its default configuration.  I see no reason, however,
why warnings should not be generated if the user, through
the advanced interface, elects to receive them.

Philip Taylor

Received on Monday, 20 November 2006 15:45:51 UTC