My process is basically this:
reset.css + 960.gs
Build site entirely in Chrome (dev channel)
Make sure it works in IE8, turn on compatibility view ensure it works there (usually a few fixed needed for IE7).
Open it up on iPad / iPhone and ensure that works (usually no issues, somtimes need to set min-width on body)
Then pop open FF 3.0 and Safari just to make sure there are no issues
(Sorry Opera folks, you don't factor into my process)
I could care less what the standards say, no customer cares, they just want to open it in the browser of their choice and see a working site.This new process makes a lot of sense as what you can use depends primarily on the browser demographics of the site and what people are willing to forgo to support IE6. It also meshes well with the release process browsers except for IE / Safari are moving to which is continuous release. You're always going to be limited in what part of the spec is 'valid' based on the browser demographics of your site, because every browser implements 80-90% of the spec.
(You can also join us. See http://whatwg.org/ for details. It's basically an open community.)
I wish Hixie would retire and get some new blood in there. He's done good work, but it's time to move on.
...and by "open" I mean anyone can join, contribute, and have a significant impact on the spec. The spec's development has been based on rational debate and research. If someone has a different vision, though, they are welcome to create a new spec, just like we did with the WHATWG when we disagreed with the W3C's vision.
You can validate using the HTML5 validator (http://html5.validator.nu/) - which presumably will be renamed to the HTML validator.
A versioned standard does not have this problem.
I realize that the response is: "validation doesn't matter". I disagree.
This is not merely about developing to the existing browser ecology. We all do that of necessity. But there has always been a somewhat-pliable boundary with versioned specs that lets you know the state of the ecology that you developed against.
That goes out the window with a living document. Many sites, such as HN and Google abandoned the DTD long ago. I do not object to this. However, I would argue that the fact that sites can choose to abandon the DTD is only enabled by the existence of a set of standards to which they choose not to bind themselves.
Abandoning a clear target is obviously not a problem today. But I am more concerned with the future: What standard are we choosing NOT to follow? I do not want to return to the chaos of the early web where we developed against browsers instead of standards. Again, I know the objections: We develop against browsers today. No, not really. You develop against browsers that have agreed to a contract of sorts to follow a coherent standard. Thus if you develop to said standard, the work required to make your site work with the current ecology is minimized. Will a "living document" lead us back to the wild west? I am not a prophet, but I am uneasy nevertheless.
I guess my main fear is that the "living document" may cease to be a target that is continuously progressing forward, and instead may become merely a description of current practices. Who, then, is minding the store?
HTML5 was a decided move away from the stagnation that had become xhtml, with a lot of new ideas. It created a cohesive goal, and moved toward that goal. Each major rev of the standard gives you an opportunity to re-assess and, if necessary, change your vector.
Also, this begs the question: whither validation? What are we validating against now?