Why is PunBB using XHTML instead of HTML?

Because Ian Hickson use twisted plot and exception to make a point he has already decided on his mind before studying the subject?

Ok fine, then why bother us with it? We know what we are doing, thank you very much, go back to bed.

Fine, make it a personal recommendation. Do not write and advertise such a personal, weightless comment, and title it ?Sending XHTML as text/html Considered Harmful?. Considered implies some kind of authority, you don't have it (yup, being paid by Google to work on HTML5 doesn't go with an ?boss hat?).

Myself, I make the opposite personal recommendation. XHTML 1 strict (sent as text/html or not, doesn't matter) is easier to validate, so it's better for newbies (people who don't validate their html/xhtml are not newbies, there are either idiots or morons). It's also compatible with xml, which make it easier to integrate with advanced tool, and probably the future (?probably? because HTML5 took a different turn, but we'll see).

Mainly, it's an opposition view thing (and the fact that the document linked by you shouldn't be put online without a strong context). He thinks (being a WHATWG member does that to you) that we should improve HTML. I think the whole thing is flawed from the start (despite its incredible achievements) and will never have proper support from UA vendor, and we should move to something else entirely (xml was a nice candidate, however I don't have preferences) that will enforce validation and proper specs.

And to be honest, this is a über-anal-geek discussion. On a basic level, for everyday use, HTML or XHTML are exactly the same thing. So it doesn't matter for 99.999999% of its user that PunBB is written in XHTML or HTML, as long as it uses strict doctype and validation.

Firefox has a very nice Vaidator extension that will pick up everything, width very useful error messages too. The link is http://users.skynet.be/mgueury/mozilla/ for those interested.  The errors you get in Opera (and Firefox) when serving the page as XHTML are only the real big bloopers (we all make typos).  So I don't agree with it being easier to validate. 

XHTML strict 1.0 sent as text/html is allowed if I remember right in the standard. XHTML 1.1 sent as text/html is considered wrong.

I think Ian Hickson's opinion does count for something, he is in the credits for helping to writing various parts of the W3C specifications  and also for helping to implement Firefox's rendering of any SGML document.  I think he knows what he is talking about.

And don't get me wrong, I like XHTML and use it when I can't care and iota about IE or older browsers.  Normally when I want to mash it in with other types of XML, that they don't support anyway.  I write my pages so it would be very simple to change to XHTML, I find it easier to come back to pages if I have closed all my <p> and <li> and most of the other stuff. I can't wait till HTML is passed out to either HTML 5 (yes that is on the way too) or XHTML 2.

Well you have half your wish of everything being xml/xslt.  XHTML is XML, even if XHTML 1 is just kind of hacked into working with a XML phaser.  I don't think xslt will become wide spread for styling web pages.  CSS is powerful enough for most people, and xslt is more complicated to write.  I can see myself useing it though, but I am not your average user.  It will take a while till we can, 15 years or so before it is supported by enough user agents and then creeps into the web developers normal tool kit.  Right now I am looking ford to seeing CSS3 working right.

I am not sure yet if I will change the doctype to HTML, I would have to change all the input and img tags.  But there is also the point that most browsers will treat it as HTML, and not only that as some tag soup since it does not have a valid HTML doctype.  Or I could just use a trick in my .htaccess to do content negotiation.