final
No this isn’t my last posting 
final is what I miss when I move from Java to C#. There are many advantages to using final in Java, not least final allows me to explicitly state that a variable cannot be change after it’s been initialized. In C# all I can use is readonly, const and sealed, and neither work within the scope of a method for a reference (other than string) 
On another topic, I like the addition of CriticalFinalizerObject in .NET 2.0 (immediately JIT compile of the Finalize method upon object construction).
Light Streamer - Real-time data (ased on the AJAX-Comet)
I’d like to second that :-). I was using final in Java very often and was always surprised that 80% or more of automatic variables within methods as well as method parameters can be declared final without any change in programming style. From this I concluded that it’s actually another instance of the 80-20 rule: 80% of variables in any program are constant. I’m only half-joiking here ;-).
What I miss even more is const modifier for methods in C++. I think it’s a super-feature when a programmer declares, and a compiler enforces, that a member method doesn’t change object’s state.
I think also that Java and C# will have to grow more features that allow declaration of certain invariants if they want to be suitable for example for compilers that optimize code for parallel computations.