Re: Problems with ? in URL and IE6 conditional statements

On 1 Feb 2008, at 16:57, Mark McDonnell wrote:
> I tried validating one of my web pages and found some errors which  
> (as far as I'm aware) are impossible to resolve.
>
> The validation link is: http://validator.w3.org/check? 
> verbose=1&uri=http%3A%2F%2Fwww.southendchildcare.org%2Fdev%252Dsite% 
> 2Fabout.asp
>
> To summarise the problems:
>
> Use of conditional statements for feeding the browser IE6 specific  
> CSS.
>
> e.g.
> <!--[if! IE 6]>
Conditional comments are a hack (although quite an elegant one) and,  
as a consequence,  they cannot be nested... although your code looks  
more like a mistake then a deliberate attempt to nest conditionals.  
Either you want to hide that line from IE6 or from all versions of IE  
that support conditional comments - so delete either line 28 or 29.

>
>
> The use of question marks in the URL (which are used to pass a  
> querystring to another page).
>
> e.g.
> <a href="myaddress.asp?id=1">my link</
That code is perfectly fine, the problems occur when you do this:

<a href=myaddress.asp?id=1>my link</a>

... and the validator gives you a very nice description of how to  
resolve the problem, including this phrase: "unless the value  
contains only lower or upper case letters in the range a-z you must  
put quotation marks around the value".

> My client works for the South-East Council and they have to have  
> the W3C validation icons displayed on the site as they are a  
> council organisation. But no matter how much I explain to them that  
> the above two problems will not affect users in any way they refuse  
> to listen to reason.

Gambling that no user will ever visit the site with a browser that  
doesn't manage to compensate for author errors isn't generally a  
great idea.

> I don't know if the W3C offers the helpful service of lets say  
> emailing me with words of the effect of "our validator is no  
> substitute for real world user testing"?

Well, it isn't - but it does make an excellent first stage of a QA  
process (since it is cheap and highlights simple errors that are easy  
to fix). (Note that this mailing list is populated largely by  
volunteers, so don't take this as official words from the W3C).

> Because other than getting a confirmation from the W3C that the  
> validator is just an automated service that can't be satisfactorily  
> relied on they are just going to keep hassling me about this.

The validator is a very reliable piece of software. While it isn't  
entirely bug free, no software is, it is extremely hard to generate  
false positives with it.

-- 
David Dorward
http://dorward.me.uk/
http://blog.dorward.me.uk/