[TxMt] Misguided use of content-type in XHTML/1.1 template

[Allan Odgaard]

On 5. Mar 2007, at 17:13, hadley wickham wrote:

> [...]
> If it's an xml file, doesn't that already imply the the content- 
> type is xml?

This is circular logic ;)

Also, we want to further specialize the type, i.e. it could be xhtml 
+xml (as in this case) or it could be rss+xml, that is what we need  
to specify.

> Isn't the content only important when you're sending things "over  
> the wire" so the program at the other end can recognise what to  
> do?  The meta tag is just a hack for html to give the page author  
> some way of overriding the content-type that the server is sending.

I don’t fully follow this. But the meta tag is not a hack, it is a  
way to specify the content type (and encoding, etc.) when the page is  
NOT sent over http, i.e. when it is loaded from a non-http server  
(like your disk drive). The meta tag never overrides the info sent by  
the server, on the contrary, it is the server which overrides the  
meta tag.

> There's a good discussion of xhtml here: http://www.hixie.ch/ 
> advocacy/xhtml.  I think this sums it up: [...]

Sure, XHTML is not useful on the net, etc. -- but what we are  
discussing here is how a correct XHTML 1.1 template should look,  
regardless of what the real-life treatment of XHTML for content sent  
over http is.

If we wanted to cater to the real-life situation, we should really  
just remove the XHTML support -- but a lot of people do want to use  
XHTML, and some may actually do it in an environment where it makes  
sense. For example some produce HTML 4 from their XHTML pages, which  
has the advantage that they can use the full XML tool suite with  
their content (when in the states prior to the final publishing step).