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

Re: Thanks a lot

From: Martin Duerst <duerst@w3.org>
Date: Mon, 10 Dec 2001 10:16:46 +0900
Message-Id: <4.2.0.58.J.20011210095519.059ea860@localhost>
To: webmaster@domovina.net, www-validator@w3.org
At 05:49 01/12/03 -0500, Frank Tiggelaar wrote:
>Over the past year we have taken great care to validate all new pages
>and pages on our site which were changed in any way. We added the small
>W3C logo to all of the pages we validated. Recently we found out that
>none of the pages which validated some time ago are validated today -
>suddenly 'character encoding' has become required.

Yes indeed. Please note that this is not a case of W3C insisting
on some ivory theory (that iso-8859-1 is the HTTP default), but
on an actual practical situation.

From: http://www.w3.org/TR/REC-html40/charset.html#h-5.2.2

 >>>>
The HTTP protocol ([RFC2616], section 3.7.1) mentions ISO-8859-1 as a
default character encoding when the "charset" parameter is absent from
the "Content-Type" header field. In practice, this recommendation has
proved useless because some servers don't allow a "charset" parameter
to be sent, and others may not be configured to send the parameter.
Therefore, user agents must not assume any default value for the
"charset" parameter.
<<<<

If user agents must not assume a default value for the charset
parameter (and all the important user agents conform to this
requirement), then why should the validator (which is supposed
to check much better than the browsers) assume a default value?


>We think this amounts to moving the goalposts during the game and our
>confidence in the w3c validation setup has completely gone.

In this case, it mainly amounts to fixing some internationalization
aspects of the validator. Such a fix was due for a long time.

Also, do you assume that if we find an error in the validator
and fix it, we should continue to claim that non-valid documents
are actually valid? Would you and everybody else be happy if
we did this?

In the case of the 'charset' parameter, this is of course not
a problem of validity or not, it's information that has to be
known *before* the actual validation can take place. Some pages
can become valid or invalid depending on the 'charset'.

Also, please note that we have done the same for DOCTYPE
declarations; before, we did some guessing that wasn't
described in any spec, now you either have a DOCTYPE, or
you don't get validated.


>Therefore we stopped validating our pages; we shall remove all 7,000
>little W3c-validated logos from our websites.

It would have been much easier to add a line or so of directives
to your Apache server setup. And that would also have improved
worldwide access to and readability of your site.


Regards,    Martin.
Received on Sunday, 9 December 2001 21:03:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:00 GMT