@Bass: This is quite a small number, especially when you consider that Microsoft usually triage a project, with x amount working on release A, and y amount working on release B., which is how they could get to Silverlight 5 in 2 years. If Microsoft invest in a project that used 20 people, any small to medium enterprise can come along and compete. Most software now is about amassing a huge amount of functionality that it becomes too costly for your competitor to create a rival product, they have to think big.
Pretty much anything written in .NET and Java can be copied easily by a competitor, which is why the software world is fraught with failed projects.
And this is the one reason that Apple shines publicly. They are very good at waiting until a few other companies flounder over a technology and work out the bugs. Then they release the same product (minus the worked out bugs) and trumpet their innovation. They get all the accolades withouth having to invest and eat any of the R&D losses.