It was last month we went to a conf about "why J2EE is not as successful as it suppose/expected to be", agenda was about real life examples, practices, different industries,

the bottom line is, stakeholders want to keep the legacy systems running (they've paid for it what'd you expect) and to match up with today's needs, they want to have the value added functionalities of enterprise  systems such as distribution  computing, controller transactions, higher security etc. But the fact is, if you, as an architect want to analyze such enhancements, you might end up with using .NET, J2EE, integration of both or even none of them. I personally don't see any value in such raw comparison of technologies without concerning the real problem statements , there are so many measurements in choosing the right framework, technology or right teams to use; which I believe you already know about them.

BTW, I think eclipse is pretty ok (note: I’m not saying it’s the best), we’ve managed to work with it in a big enough team. I agree with you as the dev tool is a good factor in speeding up the development  or how does it integrate with other dev tools, etc, but it’s not the only one. May be because I’m more into design, I guess once you have a good view of what you’re doing, you can write  the code  with ant and a text editor - ok may be I’m pushing it though