The next thing that strikes me about Silverlight is that it will rely on an already massive developer resource in the Microsoft machine, MSDN, all the existing infrastructure and tools like MSSQL, and the deep experience in content streaming with Windows Media.

To put it simply, the Silverlight juggernaut is just another display layer or front-end enhancement to an already massive developer toolset that is enabling a throng of existing developers worldwide to create Flash-like interfaces without having to learn the often quirky Flash-only language.

The final nail in the proverbial coffin for me is definitely tool sets.

I saw this article the other day, and I totally agree with his argument.  If think VisualStudio is one of the best computer programs not just one of the best programming IDE's.  Everything seems to 'just work' in VS2005 the way you expect (at worst, it's a 'right click' away).  And when you need an advanced feature, it's already there (like custom data visualization in the debugger, intellisense for almost everything, and the code snippets, 'natch).  Compare VisualStudio 2005 to Photoshop CS2 in terms of usability -- it took me years to learn the Photoshop options, and it still takes me quite some time to do mildly complex photo manipulations.

As a developer, I had always wanted Flash to succeed.  They have 98% browser penetration, and the Internet is such a good distribution model.  What developer wouldn't want his application to run "out of the box, no install" on 98% of the computers?  I tried several times to convert my application over to ActionScript, but the tools were just not mature for real applications (No debugger to speak of, 'almost' XML support, no native SOAP/WS support).  The lack of features and tools just precluded any real applications.

Wether Silverlight will kill flash or not is irrelevant to me.  Silverlight 1.1 alpha is showing lots of promise for me (it is so fast!), the key is to get it get widespread distribution.  It's a chicken and egg problem.  Until (any technology) has wide distribution, it's not worth writing software for it.  Once the technology has wide distribution, it's to late to start development.  Most of the environments I work in are totally locked down 'corporate' computers, so anything that is "already installed" is a pretty compelling environment to develop for.

Flash in the Pan: News - Software - ZDNet Australia