First off, I think there's

an51091 — Sun, 16 Mar 2008 12:16:38 -0400

First off, I think there's an important distinction in RIAs between public internet applications and enterprise applications. Things like rich design support, Silverlight's Deep Zoom feature, and related eye candy are primarily for public websites building internet applications, not for enterprise apps. For internet apps, I think it's a wide open question. Flash has enormous traction here, but primarily as a design tool, not as a programming tool (yet). Meanwhile, Silverlight has very limited traction here at all. Thus far, the winner is AJAX, mostly by default. Whether either platform can become predominant for rich public website applications is, I think, a wide open question, and I think it's just too early to call.

For enterprise apps, which is what you primarily address, the importnat attributes are very different. As long as your application isn't actively ugly (and maybe even if it is), design just doesn't play that important a role. The reality of enterprise applications is that the vast majority of them are written using two platforms: Java, and .Net. The enterprise world is slowly opening up to Python and Ruby, and there are still bastions of C++ and older technologies, but Java and .Net are clearly the most important platforms.

Silverlight's #1 factor that assures it's success in the enterprise is that if you're already developing for .Net, moving to Silverlight is going to be a no-brainer. That doesn't mean it will beat Flex, but the reality is that even with Beta 1 (and some pretty basic missing features, like combo boxes), it's already easy to tell that Silverlight 2.0 is going to be a better platform for developing enterprise apps than asp.net, which is it's real competitor. And because you can develop Silverlight seamlessly in Visual Studio within the same solution as your server-side project, I can't imagine .Net developers defecting to something like Flex for a front-end, because the integration will simply be too much trouble compared to Silverlight. In this sense, I think the sheer breadth of server side tools that a client side tool integrates with is far less important than how tight that integration really is.

Of course, the fact that Silverlight works so seamlessly with .Net is great for .Net shops, but that's exactly why it will never get any traction in Java shops. Let's be realistic: if you're developing in Java, using Eclipse, NetBeans, IntelliJ, or whatever, moving your client-side development over to Visual Studio is about as likely as switching to Emacs. So the only realistic candidates for Java developers are solutions that will integrate solidly with their existing toolchain. The obvious possibilities for Java-based RIAs include Flex and JavaFX. Flex obviously has a head start, but it involves developers moving to a different language, set of libraries, etc. for client-side development than they use on the server. The real question for Flex is whether Adobe can convince a preponderance of Java developers to make this leap instead of waiting for JavaFX or sticking with some sort of AJAX framework. I don't keep close enough tabs on the Java community to know how this is going.

I set aside dynamic languages like Ruby and Python before because their enterprise penetration is currently so low. However, this is actually one very interesting wild card going forward. Both Silverlight and Flex have pledged to allow client-side development to occur in a variety of dynamic languages in the future, and it will be interesting to see how this plays out in these language communities. The current beta of Silverlight already allows Ruby and Python developers to experiment with developing client-side applications in their language of choice (and supposedly PHP support is coming as well). In theory, this should allow clean interoperation with server-side applications in these languages using web services, though in practice, there's really no build-out of tool support for this yet because everything is so new. Will the growing pool of dynamic language developers find Silverlight appealing as a rich client solution (despite traditional indifference to Microsoft technologies), will they wait for Flex support, or will they stay with their current AJAX-based solutions? Offering this is a great idea on Microsoft's part, but I honestly have no idea whether they will be able to gain traction.

.NET Zone - Comments for "What Does it Take to Be a Good RIA Platform?"

First off, I think there's