Herenvardo - 4:58 am on Aug 25, 2008 (gmt 0)

Well... let's start with this. I wanted the opening post to be completely neutral, and only provide factual data (and the key references) to provide a good starting point for the debate. So I'll start posting my own opinions from this post onwards. But before that, some more bits of factual data, in reply to your post:

If they make the newest browsers of today (IE 7 & 8, FF3, etc) compatible with XHTML 2 and release 5 to 10 years from now, it may very well be a viable option.

It would be good to know which release dates are being targeted, wouldn't it? ;)
According to the WG charter at W3's site, the timeframe for XHTML 2 was defined as

Following its approval by W3C Members, this group will commence in August 2002 and will terminate in August 2004.

It doesn't look too acurate. However, in March 2007 the XHTML 2 working group (which is the previously known as HTML WG) was chartered (that's the point when W3C embraced HTML5 activity, hence having to redefine the Working Groups), and a new timeline was set:
First public WD: August 2002 (this was defined as the date when the draft had been published, so it is obviously achieved)
LC: September 2007: it seems that they are almost a year late already
CR: February 2008: already late
PR: June 2008: also already late
Rec: September 2008: hurry, hurry! you can still make it! :P
For the HTML5 charter, they have used a list layout instead of fancy tables, which means that I can quote it properly here:
You can skip all this part unless you are interested in what are the Milestones for an W3C spec and how the whole process works.

1. 2007-05 HTML5 and Web Forms 2.0 specs adopted as basis for review 2. 2007-11 HTML Design Principles First Public Working Draft 3. 2008-02 HTML5 First Public Working Draft 4. 2009-03* HTML5 Last Call Working Draft * editor's estimate is 2009-10 5. 2009-06 HTML5 Candidate Recommendation 6. 2010-06 HTML5 Proposed Recommendation 7. 2010-09 HTML5 Recommendation

The third point is already late: there is an "Editor's draft" publicly available, but it's not considered a Public Working Draft (an Editor's draft would normally be only available for WG members, but this WG has "public proceedings", so everything is available to the public). Also, if you take a look at the draft (the link in the opening post, since it is the latest draft) you will understand why this wouldn't normally be "published" :P
The 4th point is unlikely to be meet on time (a better chance if we take the editor's estimate instead, which is a more educated guess based on the pace at which feedback is dealt with). Once you read the WHATWG's FAQ [wiki.whatwg.org] (and my explanations below about how the W3 works), you will find out that the points 5 and 6 are going to take very long, and the WHATWG's estimate for HTML5 becoming an W3 recommendation is, as stated in the FAQ, "in the year 2022 or later".

About how the W3 works (and about all these flavors of specs):
[b]Editor's Working Draft: that's the working draft the WG works with on an every-day basis. When it becomes something "human-readable", a Public WD is released.
Public Working Draft: A date is defined for the first, because there will surely be a first public draft, but there might be only one, or there might be dozens. The main goal of the public draft is to draw feedback from the different interested parties (although the HTML5 WG is doing this directly from the Editor's draft, which is also public in this case, so this one doesn't have too much meaning). Feedback is normally translated into "Issues" and issues are dealt with by the group. When all the issues are dealt with, the next public draft is released, and so on until a draft arises no issues (due to the openness of the HTML5 process, this is being done in a much more dynamic way, but the concept of issues is similar: once there is a draft that has no issues and no new issues are opened in a given timespan, the spec advances to Last Call).
Last Call WD: Also known as "Call for implementation", the release of the LC denotes some stability of the specification and invites interested vendors to implement it. At this point only implementation feedback is gathered. If this feedback translates to new issues, a new LC adressing such issues is published, and so on until a LC "survives" long enough (I'm not sure how long it takes) without new issues, it moves forward to CR. In the case of HTML5, since major browser vendors are already implementing the most "stable" parts of the spec on the run and providing feedback, the LC stage should be quite short for this case.
Candidate Recommendation: Basically, on this stage test suites must be made available, so implementations can be tested (I'm not sure if test suites need to be ready at some early stage, however). Once there are at least two public, general use implementations (ie: no betas, experimental builds, and other unusual stuff are counted, only public "final" releases) that are valid according to the test suites, then the spec takes a step forward in the process. This is likely the point where HTML5 will take longer: according to the FAQ "20,000 tests for the whole spec would probably be a conservative estimate". Actually, the same FAQ states that they aim for reaching CR by 2012, which means just 4 years from now; and then 10 more years only for the "testing" part.
Proposed Recommendation: At this point, the only thing to be done is to vote the spec within the W3C. If the voting is successful, then the spec becomes a Recommendation. If it's voted against... then I don't know what it happens :S .
Recommendation: At this point, the specification is an "oficial" W3C standard, and in theory it may be relied upon by content authors and tool (including authoring tools, browsers, and any other relevant software) vendors.
Here ends the boring part.
In summary, the specification should become stable by 2012, and become an official standard by 2022. However some parts will be available for developers to use even before that: for example, major browsers already implement the <canvas> element; and many features of HTML5 can be emulated through JavaScript to achieve a "smooth" adoption.
So, that's everything you may need to know about the dates. If you want to be aware of updates on these estimate dates, keep checking the relevant links.

Now, my opinion:
I think that, comparing the documents right away, XHTML2 is currently better: it provides a cleaner content model, draconian error handling is not an issue for me (I author compliant and well-formed documents), and it provides some improvements that I really like, such as the sectioning model, or the complete separation of presentation, structure, and semantics. For example, if you want to have a list for navigation (and be recognized as a navigational element by supporting UAs), in XHTML2 you'd use the <nl> element, probably with a role="navigation" attribute. In HTML5 you would need to use a <nav> element, which defines both the "navigation" semantics and sectioning structure: if it is navigation, then it will be forced to be a section. If you indeed wanted the navigation to be a section, XHTML2 allows you to define that as
<section role="navigation">
<nl>
...
</nl>
</section>.
Another advantage of the XHTML2's sectioning model is the <h> section header element, which is equivalent to the <h1..6> appropriate for the current level of <section> nesting. This becomes really useful when maintaining a dynamic site: you added a header and now you need to uptade all the <h3> and deeper tags in who knows how many other scripts and includes.

Of course, browser vendors' support for HTML5 is also an important factor, but it will probably get eclipsed by an even more important factor in favor of HTML5: although the XHTML2 spec is currently better when analized by its own merits (ie: without counting that browser vendors have already make their choice, or that it doesn't deal with the existing legacy of tag-soup on the web), it is already "closed" (it won't be changed at all except for the redundant appendices and editorial corrections), while HTML5 is more open than any other spec in the W3C's history: the spec is being built by the ~900 subscribers of the mailing list, with the WG members acting mostly as moderators. And I can state first-hand that the feedback provided through the lists is properly dealt with. This doesn't mean that every feature request will automatically become a HTML5 feature, but that the Editor and other group members will review it and consider it based on its technical value (ie: if it solves a real problem, it will most likely in; but if you ask for a <blink> tag without even exposing a reason you will most likely be ignored).

So all in all, I feel that currently XHTML2 is, technically speaking, better in many aspects; but HTML5 has the potential to become even much better than that. Adding to it that it has most of support, it is most probably going to "win" for good.

Regards,
Herenvardo