Everybody is talking about why you “should” serve your XTML pages as “application/xml” or something similar. I really think that they are missing the point.
In my opinion the most important reason for promoting and using XHTML is not that it is XML and therefore inherits all the programming possibilities. The main reason is the coding practice that follows from its use.
Writing user-friendly (read developer friendly) code is the most important issue here. And not the fact that you can write valid code that adheres to some scheme or DTD.
Getting your code to validate is just as fun and enlightening as beating your head against the wall.
Everybody relax; getting your pages to validate is not the main goal. It is a means to an end.
But before we get there, we should learn to write easy to maintain code. This will make things easier to debug in search for validation errors if we want to get validated code.
Coding your pages in XHTML “must” mean that you write semantic code and try to separate content from presentation. Most people tend to apply a few simple rules (close your FONT tags!!) to generate a valid impression of XHTML. But unfortunately they are completely missing the point. (Though this will get them the goodies that come from using XML syntax.)
XHTML is rich and can really add meaning to a document. The presentation from Tantek emphasizes this and shows that the tags are really designed to tell the developer the contents purpose inside an element. (Remember that XML was designed for humans and machines.)
It is the same thing when you write code in PHP, VB or Java. Most of the time you use a naming convention for your variables and functions. Just to make it a bit more easy to read and understand.
Ever read some code that uses function names like: function do_aRt() or variable names like var aRt_01 ? Doesn’t make sense, does it?
That is why they invented the Hungarian notation, or camelcase for writing more developer friendly class names and variables.
It makes things much easier if you want to review your code after a year.
Same thing counts for XHTML. If you use descriptive elements, things will be so much easier to read and to maintain your code (assuming that you read your code).
The following is just an example to make things a bit clearer.
Suppose you are hired to do the layout for a Chinese or Arabic website. You get to write the HTML templates and the CSS files, based on the existing website.
The first view on the code shows me tagsoup, with several nested tables and a few DIV’s and FONT tags. Where to start?
If the previous developer used tables for layout and never added the slightest indication of what the function of the content was, you are lost.
Back to the owners and request a detailed wireframe in English!
Start with really thinking about the content that you are trying to display and maybe 1 of the 91 tags could add some meaning to that. The Markup reference from Mozilla really does a good job with adding good semantics to existing tags. And of course read the SimpleQuiz from Dan Cederholm.