Flash/Flex/Air are making some in-roads, and the desktop application as we know it is getting it on the chin! It's an inevitable collision. Through the adobe technologies people can leverage the connectivity and a rich user experience. The browser is still required by most of this new generation of applications, but speaking with adobe people the paradigm is shifting away from the browser and "urls".
Nothing new here right? After all we've been speaking about this since Java came about in 1995. What is different, well I've been playing with Blist a database application, like access or Filemaker, or Corel Paradox(which I worked on back in the days). There is another one called joggle a photo management package(I haven't had the chance to try this one out).
What's common about those two apps and many others- They are slick, powerful and with no compromises on the user experience. Java on the client never delivered on that. The Adobe trio does. This is a wake up call for us all working in the shrink wrap space, up and coming smarty pants entrepreneurs in a basement are coming out of nowhere with offerings that are right up there in terms of value and experience.
The downfall of the platform from what I can gather around the web is that when it's time to do computations it's not up to snuff, compared to java, which in turn is not up to snuff when compared to C++.
I think we need to look seriously in Adobe's offering because it's got some powerful tools. The platform has momentum, offers the ability to develop fast. Last characteristic, but not least, it's a great equalizer the barrier of entry to complex interactive and connected applications is lower than before. I was told once - Never underestimate the market appetite for tools that will make them look smart and talented ie make them shine!Too often in software engineering we think of building products using tools for real men, or technologically elegant, and we forget that time to market, the ability to iterate rapidly, and provide a compelling user experience is what really matters. We have to open our minds to tools that from the first look could be perceived as unworthy, or difficult to integrate with the existing code base.