W3C home > Mailing lists > Public > www-validator@w3.org > May 2009

Re: code too often not accepted

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Sun, 3 May 2009 23:30:13 +0300
Message-ID: <9392B9EED515432C8BE1F652D0AAAFD3@JukanPC>
To: "Joop Nijenhuis" <joop.nijenhuis@hccnet.nl>, <www-validator@w3.org>
Joop Nijenhuis wrote:

> Trying hard to get html-pages correct coded, but I have a hard time.
> Your program keeps on bashing and at the moment I think its all to do
> with interpreting and translation errors.

It seems that you do not know what markup validation is. That's all too 
common, but it tends to lead to problems when trying to use a markup 
validator. Please check this introduction of mine:
http://www.cs.tut.fi/~jkorpela/html/validation.html

> It might be a big help if
> you can go from the errpr through the rules on which you based YOUR
> program.

The validator works, or at least should work, by the rules that YOU specify 
in your markup, in the DOCTYPE declaration.

> Saying that its not allowed and that something is missing or
> wrong while other often nationalised "papers" do say its allowed
> keeps us going nowhere.

It does not lead you anywhere unless you learn what validation is and what 
the validator is doing, or should be doing, and unless you are willing to 
check the syntactic definitions as needed.

> to give you an example;
> <start>
>
>   1.  Error  Line 52, Column 145: Attribute "SCROLLING" is not a valid
> attribute. Did you mean "scrolling"?
>
>      ?load()" BGCOLOR="#EFEBAD" scrolling="no">

This is unfortunately an example of grossly misguided second-guessing by the 
validator, and a recent "development" if I remember correctly. Just read the 
first sentence with the addition "for this element" and ignore the second 
one.

The validator tries to be helpful here but fails miserably. Its authors know 
that people often forget that case of letters is significant in XHTML. So 
they souped up that "Did you mean 'scrolling'?" thing, which is all wrong on 
two accounts:
a) The markup wouldn't become valid if the attribute name spelling were 
changed that way - but the "helpful note" strongly suggests otherwise.
b) The message is also given when an HTML 4.01 document type is used, 
despite the fact that in HTML 4.01, case of letters is not significant in 
attribute names, i.e. scrolling and SCROLLING (and ScrOLLing) are treated as 
the same name.

> May be you can explaine what to do here? I'm out of options.

Check the document type definition that your document purports to comply 
with, or maybe its prose description in the HTML 4.01 specification.

> On the
> Dutch help pages you have to use scroll in the body argument, like
> <BODY SCROLL="waarde">. with this I started, but wasn't accepted.

It's just nonsense. Even if it weren't, it would not comply with the 
document type you use.

As the validator's explanation (and explanations you've got on this list) 
says, you're supposed to use CSS to prevent scrolling (which is almost 
always a bad idea for the <body> element - test with a stamp-size browser 
window and you'll see why). Strictly speaking, this is completely external 
to validation. The validator simply says that the attribute SCROLLING is not 
allowed for the BODY element, without saying anything about what it might 
mean if it were allowed or what you should use instead etc.

Don't expect the validator to say anything about things that are not markup 
syntax issues. If it does say something about them, don't expect it to be 
right.

-- 
Yucca, http://www.cs.tut.fi/~jkorpela/ 
Received on Sunday, 3 May 2009 20:31:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:35 GMT