For a long time I have been looking for ways to make in easier in C# to manipulate datastructures, as this is most of the code you write (aside from checking business rules) in Enterprise apps.
Things that would be especially usefull, is that when working with datasets or XML documents, the compiler can acutally help you check for correctness data manipulation code, for instance check the syntax of your XPath expressions etc. Normally they are just a string and the compiler leaves them alone.
Another features that I've wanted is to perform SQL like operations on client side datastructures, such as datasets.
Microsoft Research has been working (and is still working) on C Omega to solve these problems, by extending the C# language with code constructs specifically target at working with data structures. I think it looks very promising and that it will make it easier writing service oriented applications, because it will make interacting with Xml style datastructures easier. Intergrate this language with Indigo, and we will have something really cool and productive...
Read all about it at MS Research and MSDN.