W3C home > Mailing lists > Public > www-validator@w3.org > April 2007

Re: Problem with validator

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Fri, 06 Apr 2007 20:47:49 +0200
To: www-validator@w3.org
Message-ID: <461695D5.6DFB@xyzzy.claranet.de>

David Dorward wrote:

>>> The end tag for the P element is implied when the parser reaches
>>> the start tag for the FIELDSET element. So, no, there is no
>>> accompanying start tag.

>> One of the many reasons why using HTML 4.01 instead of XHTML 1.0
>> in manually edited documents is a dubious plan.
 
> Not really. Fieldset isn't allowed inside paragraphs in XHTML 1.0
> either. The only difference is that the error message emitted by
> the validator is confusing in a different way

It won't bark at the </p> without first reporting that something
is wrong at the <fieldset>.  Those implicit tags in HTML are hell
for ordinary folks not knowing the DTDs and the SGML rules by heart
- maybe for everybody minus Jukka and you.  

> i.e. suggesting that an object/ins/del/etc element is the solution

At least the confusing message is at the right place for XHTML 1.0.

> The problem comes from authors using the validator as a substitute
> for not knowing the language constructs they are using.

When I check my own XHTML 1.0 documents it's always clear what went
wrong, mainly typos, ":" instead of ";" is my favourite.  Sometimes
I'm forced to check HTML pages too broken for my old browser, and the
first time I stumbled over the "implicit </p>" issue it had a "high 
astonishing factor".  Of course I wouldn't try to start a <pre> "in"
a <p> on my own pages, but it was on a page written by somebody else,
the result of a template in a template which used to work perfectly
until some "redesign" effort put the stuff as is in some table within
table within table layout... <shudder />

Of course it "worked" with all popular browsers, but not my dinosaur.

> Using XHTML 1.0 in manually edited documents is even more dubious,

Editing XHTML 1.0 manually is fun, the strict XML rules make it hard
to  create incorrectly nested tags, nothing is implicit.

> especially if the document is to be served as text/html.

Of course it's served as text/html, it's designed to work with "any"
browser, almost a decade ago.  Learning appendix C by heart was easy,
my browser made sure that I did.

Maybe Web servers could send it depending on what the client likes
better, but for that they'd need to know which *.htm or *.html files
actually are XHTML.  For simple "Web space" offers configuring this
by the end user isn't possible (and admittedly I doubt that they'd
get it right if they're allowed to try it anyway).

Frank
Received on Friday, 6 April 2007 18:50:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:24 GMT