Re: [xml-dev] 2007 Predictions

["Kurt Cagle" <kurt.cagle@gmail.com>]

Len,

I'll disagree with you on this one:

2007 is the year that the HTML/CSS browser as the ultimate wrapper for web
content begins to see serious competition from standalone applications where
market revenue recognition predictions are impacted by the complexities of
relying on plug-ins.  The WHATWG vs W3C shootouts indicate that the constant
massaging of the HTML frameworks to get small competitive advantages from
loss-leaders affects the emergence of other markets in the plugins that do
not actually need other than standard network support.  In short, the HTML
browser is finally seen as the market pig that it is.

I think the "HTML" browser is definitely creaky, but I think the mixed mode XHTML browser is also just
really coming online. Mozilla is making serious in-roads into the enterprise level now (at least from anecdotal
evidence, though I don't have formal numbers) and will likely continue to do so. Mozilla is also creating a
vacuum effect - its becoming the browser to emulate (or outdo) in terms of core technologies, and the Mozilla
implementation is quietly replacing the IE implementation as the canonical standard that's forcing other vendors
to create feature equivalents. This effect has already forced both Opera and Safari to migrate to an XML aware
architecture, and the continued churn in the AJAX world will likely end up pushing at least one of the stacks (as a
guess, I'd say DOJO) into a C++ implementation with JS exposure.

XAML, FLEX, OpenLaszlo, Boxely and other XML frameworks are opening up, and I think that the XML-based stand-alone apps MAY be on their way in, but I honestly don't see that market opening up appreciably until 2008 at the earliest. There is also still very much a deployment issue, and the fact that people are increasingly spending more and more time within the browser context that will be a strong inertial factor to overcome.

My own feeling is that the XML client binding framework likely will be one of THE key XML stories of 200, the others being XQuery and XSLT2, the latter of which I expect to see some SERIOUS traction on. I talked recently with an IT Director for Lexus-Nexus, and he indicated that they are increasingly using XSLT not so much as a presentation layer generator but as a router, bypassing a lot of commercial tools to do so. I disagree with Jon Bosak on that one point ... I think that with the simplified transformations that XSLT2 opens up, and the consistent extensions mechanism, that XSLT will likely end up becoming much more commonly used as nodal routers evaluating XML business logic instead of specialized web services doing the same (especially since such XSLT actually works very nicely with a messaging queue). Make XSLT callable from within XQueries (see eXist for a wonderful implementation, especially with Saxon 8 doing duty as the transformer) makes for a considerably simpler yet more powerful web architecture than lots of code monkeys writing linear ASP/JSP scripts.

-- Kurt