Peter, you hit on a good point...

Using a bounding table in your skin is something a designer could correct (when the time comes) fairly easily.  This is a big benefit of using a CMS in the first place. Poor module design, on the other hand, is a more enduring issue.  In a lax environment developers are allowed to make poor (dated?) judgment calls and get away with it. 

I would love to see these changes:

1. DNN corp make standards compliance a factor in certifying content for Marketplace. 
   
   Joe, as the author of the guidelines, you are specifically empowered to steer the DNN ship toward compliance.  Perhaps include a test of the skin/module at WC3 and report the number of critical or low level errors?  You check for Standard Portal Styles... how about Standard Markup
2. Take on the development of a framework like YUI to make compliant development a shorter learning curve.
3. Expand the word "superior" in the mission statement to include specifics like compliance, usability and accessibility performance measures.  IE, clearly state a set of principals or values, not just procedures, to guide the development of the framework.  (my apologies if I missed this content somewhere)

( I think we lost Brent ;) )

Such comments are unproductive.  Just because I may not CHOOSE to go with a "Pure CSS" layout doesn't mean I don't know how.  I am sure that goes for quite a few folks that CHOOSE not to use a "Pure CSS" approach to design.  It is not always ignorance.  That is a poor assumption on your part and insulting to say the least. 

A mindset saying, "I do things this way, so I am better than you and you're wrong," does not help anyone understand the why and how of this topic.  We would all be better served if you would tell us your experiences as to why you use CSS and why you don't do things any other way. 

Even though I am one of the people that use a hybrid approach to HTML design and layout, I am certainly interested in knowing your unique experience and knowledge in this area. 

OK I'll calm down and come up with a thoughtful response instead, then.

I choose CSS / XHTML "pure" coding because it is the best possible way to completely separate your markup from your presentation (even though some would argue that markup is presentation).

But when you talk in terms of "un-styled" vs "styled" then semantic markup (without tables) plus CSS is the best way to go.

My favorite test of how effective your markup is is to disable your stylesheets and see what your content looks like in pure markup. You should still be able to make sense of the content and see headers and paragraphs and lists. This lets you know that your web site can effectively communicate with other machines, such as search engine robots, screen readers and mobile devices. That is the direction the web is heading.

Just compare a complete semantic / css layout with a table layout both with CSS turned off. The table layout will not conform to the screen size as readily as a semanticly marked up html page will.

Every thing that is readable to the human eye should be readable to the computer, which means using CSS to it's fullest extent to style elements so that there is a text version that is hidden when CSS is enabled. For example, using text-indent and background-image if you have an image with non-standard fonts in it:

<h1 class="hide" id="logo">Company Name</h1>

h1#logo{

width:100px;

height:100px;

background:url(images/logo.gif) no-repeat;

}

.hide{

display:block;

text-indent:-99999px;

outline:0;

}

----

This approach works a million times better than

<table cellspacing="0" cellpadding="0" border="0" width="100%">

<tr>

<td><img src="images/logo.gif" alt="company name" /></td>

</tr>

</table>

Because when CSS is disabled you still see a semantic reprsentation of that data: The H1 tag.

Now when it comes to layouts, it is just a matter of reading up on more advanced techniques to create effects that were previously generated by frames and tables. www.alistapart.com/articles/holygrail

Read up on that article for advanced technique in creating CSS layouts that look great in any browser and work in any CMS. Next, read up on Semantic HTML for why it's important to use markup that is intended to represent its content.

Tables do not do a very good job at representing their content unless their content is tabular data.

The reason I bring this up is because I'm pretty tired of watching the .NET community completely miss the point of standards, accessible, usable and semantic HTML. Here's a fantastic article on 456 Berea Street on the problem with programmers not fully learning the entire scope of their trade.

Now maybe my comments were innappropriate but I find it unbelievable that in 2008 people still wonder about when to use tables and when to use DIVs and by DIVs what they're really searching for(hopefully) is Semantic HTML and not just replacing the table with the DIV.