I gave two new talks at SD 2007 West today, Web Forms 2.0 and What's New in XML in Java 5 and 6. The Web Forms 2 talk was especially fun. It covered both Web Forms 2.0 and HTML 5. I think there's a lot of pent-up demand for new HTML, which in many ways hasn't really changed since 1999. HTML has one of the better models for forward and backwards compatibility in tech. This in large part drove the Web from about 1992-1997 or so. Sadly no one has taken advantage of that to push HTML forward for a long time.
The Web is actually in better shape to introduce new markup than it has been
before. Well-formedness is part of that. Separation of syntax from semantics
is critical. Sadly the WhatWG is actively hostile
to well-formedness and XML. They are mired in the outdated 1980s model that
requires all semantics to be defined up front.
That's crippling them in several ways. The second piece we need
to really let the Web jump forward is a simple change in the browsers.
They need to start allowing CSS to style and script
unrecognized elements, not just ones they
know about in advance.
The key is to let the browsers
handle what they understand,
and provide stylesheets and scripts for the rest.
Currently they simply throw away the pieces they don't understand.
That's better than rejecting the document
completely, but it's not as good as allowing stylesheets and scripts to woork on those parts.
This would enable us to start using nav, m,
header, footer, meter,
and many other proposed elements now without first
revving the browsers. Of course, this really does require
end-tags on everything. However, it's much more powerful and simpler.
The WhatWG approach of specifying everything in excruciating detail
will take far longer and achieve less.
The alternative is to serve nothing XML to the client and let the stylesheets do all the work. However JavaScript doesn't work with this, and we go from some predefined semantics to none.
I've also posted the notes from today's Effective XML, class at Software Development 2007 West.
