David Andersson has an article at Digital Web about XHTML2 and HTML5. I'm firmly in the XHTML2 camp - the features are more what I want, regardless of browser support. If a whole load of developers stood up and said "we support XHTML", the browsers might fall in line and pay attention. For too long, the Web has been all about browsers, and people forget that people do things with Web data. 
For too long, the discussion of Web standards has been far too conservative and has lacked any vision. 
What we need to have is a dual model approach - instead of having an HTML WG and an XHTML WG, the W3C ought to combine the two and have a branch to work on fixing up what we're doing already and a 'future' WG. 
The 'current' WG would fix up HTML 4/XHTML 1 to clean up problems with them, and work with well-established technology like CSS - and basically try and harangue the browser makers in to supporting (X)HTML and CSS properly. The 'future' WG would work on extending that work towards new XML based constructs - SVG, RDF etc. 
That way, the 'current' WG could operate on what's practical (ie. what web designers are bitching about) and the 'future' WG could work on stuff that's just plain cool (ie. what web designers would like to dream about if they weren't bitching about other stuff). 
As the stuff that the 'future' WG does gets more and more practical, it could be neatly transitioned in to the 'current' WG. This model works quite neatly with software development - stable and unstable branches. Of course, both the HTML future WG and the current WG would have to be open processes with public oversight. 
The 'current' group should focus on browser support, accessibility, usability, inter-operability and well-specified standards - on paving the cowpaths, on keeping it simple, on building consensus. The 'future' WG could then focus on building the web of data, rethinking old assumptions, doing things differently and writing standards to fit the niche cases. The 'future' group wouldn't be paving cowpaths, it'd be dreaming of Platonic highways, coming up with grand designs that maybe aren't practical. 
The synthesis of current and future would get us where we need to go - having three competing standards bodies (HTML WG, XHTML WG and WhatWG) pushing different standards based on different paradigms (SGML/HTML 4 and XML) probably isn't. Just an idea. 
