Based on how fast I see my peers picking up JSON, I think the
shortcomings of the DOM API have been underestimated. At the same time
I think having JSON specs for RSS, RPC, etc is going to start to seem
crazy to everyone. Something like e4j could take off in a big way.
It's a chicken and egg problem, and I suspect that JSON is going to start losing steam as it pushes against the lower boundaries of what XML really works well for. The biggest cost with XML right now is parsing to and serializing from DOM - if you create a lightweight language such as E4X, these costs drop, perhaps not quite so much as for JSON, but certainly within the same general ballpack. E4X also integrates well with JavaScript (good for small-lightweight transformations for which XSLT is overkill), and E4X/JSON converters are fairly trivial to write, for those who absolutely MUST work purely with objects. In the end, most of the compelling arguments for JSON over XML have to do with the complexity involved in programming with DOM. I WOULD like to see a Mozilla implementation of XPath in JSON (it's in the original ECMA standard), but beyond that I agree with you on the rise of such efforts as e4j, e4x, xj and Linq as being a major story for 2007.
In 2007 the combination of increased browser support, increased literacy
and a technical and social climate that encourages it could lead to an
explosion of XML use in the browser: A revolutionary non-revolution
much like Ajax.
Ajax has been driven (with apologies to those on this list) largely by script kiddies until comparatively recently - young hot shot web-devs who are out to prove that they can parse this tune in five notes or less. As Ajax moves into the enterprise, and starts having to deal with corporate data streams, the limitations of JSON are going to become increasingly apparent, and there WILL be a pushback for AJAX to start incorporating better XML support.
There are already a lot of things you can do with mozilla that you just
can't do with IE. 2007 could see Mozilla-only functionality being part
of the next killer app. That's an internet with IE7 left behind.
Gotta be keeping certain people awake at night. ;-)
One phenomenon that I'm seeing with IE7 that could be a very bad sign is the number of people who are trying DESPERATELY to get it off their system. These aren't Microsoft bashers ... they are the Aunt Mildreds of the world who are suddenly faced with a browser that is acting very flaky and causing problems elsewhere in the system. This may not actually be IE's fault, but the perception is rising that it is a regression from IE6. I've heard that IE8 is now circulating as an alpha up in Redmond.
Mozilla is facing growing pains of its own, but I think some of the problems that it faced earlier have largely been resolved. I would suspect that there are more than a few pilot programs right now being built around Mozilla, and getting SVG support wrapped, the XForms extension solid enough to integrate and things like the local data-store documented and put into play will continue to push development there. I'm also paying more attention to Apple's WebKit, as I suspect it may end up providing an interesting alternative that will nonetheless tend to solidify the "Mozilla" paradigm, if the APIs I'm seeing are any indication.
> 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.
I see the browser as the OS for these apps. Is that too cliche?
Nope - OpenLaszlo is already moving that way, creating a DHTML layer as an output format for their XML component-oriented input. Boxely is interesting, but I have some concerns about the license. XAML is not taking off in the way that MS expected, though it is attracting a developer-base, but I suspect that XAML will continue to be a secondary technology until such time as it is plays nicely with the IE browser - or a Linux version of it exists (and yes, I know that's on mono's radar).
