Which makes it even more extremely suspicious when a supposedly independent third party releases performance tests on the product. Of course, they had to make thinly-veiled comparisons to BEA and IBM. Somehow, I think there will be some serious legal action on that front.
Besides, all the performance tuning of .Net was done by Microsoft employees! How can you consider that a third party comparison?
Ultimately, the suitability of any solution depends on far more factors than performance alone. Most large scale companies will not let Microsoft into the enterprise space, even over their dead bodies.
I have worked on many Java implementation, and find that it can solve some situations admirably, especially those with distributed data or code objects. But then, a lot of work can be done in C-Shell languages and PL/SQL, and have no need for Java whatsoever.
The independent advisor will look at the complete situation, analyze ALL costs as best possible, and make the most informed decision regardless of personal ideology. It can be near impossible at times to overcome the marketing hype on all sides and recommend the best solution. If you want a REAL political fight, try to deal with comparisons of DB/2 to Oracle, and throw in a mainframe option to boot. Then you will see sparks fly!