On Nov 27, 2008, at 5:32 AM, Michael Ludwig wrote:

>
> Robert Koberg schrieb:
>> I would hate to see a schema that validates a CSS based XHTML page.
>> Even worse, I would hate to maintain it.
>
> Why would you hate to see, or have to maintain such a schema? Is it  
> the
> additional tedium you find uninspiring? Or do you think it is a bad
> application, or even an abuse of XML Schema? If it is the latter,  
> could
> you explain why you think this is bad?

Well, I guess I should have said that I don't even know how to write  
an XML Schema that will validate a CSS based XHTML page. For example,  
a div element with a particular ID (a unique NCName) must have a child  
div with another particular ID, but also allow other divs/tables/ol/ 
ul. The another div with another ID must have a child that contains a  
certain class, but also allow other divs/tables/ol/ul, etc...

>
>
>> However, it would be pretty easy to use an XSL to validate your page.
>> You could base the XSL off of your CSS rules. For example, you might
>> have a CSS containing:
>>
>> #narrow-col #nav { blah... }
>
>> <xsl:template match="*" mode="validate-nav">
>>  <p>
>>    Has Nav:
>>    <xsl:value-of select="exists(key('ident',
>> 'narrow-col')/*[@id='nav'])"/>
>>  </p>
>> </xsl:template>
>
> That's a good idea. Strangely, it hasn't occurred to me I could use  
> XSLT
> for the job. It doesn't seem to have a reputation for being a  
> validator
> tool, yet.

Well, there is schematron (http://en.wikipedia.org/wiki/Schematron).

Like you mentioned, you can just as easily use HttpUnit. You might  
also want to look at WebTest (which uses HttpUnit) http://webtest.canoo.com/webtest/manual/WebTestHome.html 
  -- with either you can use XPath to find what you need.

However, I find using XSL to be easiest and most flexible.

best,
-Rob

>
>
> Thank you very much!
>
> Michael Ludwig
>