More to the point, C# has integrated an excellent map-reduce framework in the guise of (P)LINQ. Java still doesn't have anything analogous to that, nor does it plan to, if the latest roadmaps are any guide.
I'm not certain that Java really needs this. Developers that wan't a more functional style should consider making a transition to Scala or Clojure.
Scala can evolve at a much faster pace than Java as well, given that the core language is tiny and most of the language constructs are implemented as libraries.
He didn't say they weren't useful, just that Java the Language doesn't need it, due to the fact other languages that have interoperability with Java on the JVM have these features already. .NET isn't as language rich as the JVM currently unless I missed some major changes.
I get that you can call across to another language that has more functional features. But that's not the same as having little bits of it right there. e.g. when you want to find the latest item in a list, instead of a loop, you do:
and carry on. It's great. You don't tend to have that fine granularity when working across languages.
I don't know if you missed F# or not, but if you want a fully functional language in the ML / Ocaml lineage on .Net, you can write some code in that and interoperate.
For example, in the type system we do not have separation between value and reference types and nullability of types. This may sound a little wonky or a little technical, but in C# reference types can be null, such as strings, but value types cannot be null. It sure would be nice to have had non-nullable reference types, so you could declare that ‘this string can never be null, and I want you compiler to check that I can never hit a null pointer here’.
50% of the bugs that people run into today, coding with C# in our platform, and the same is true of Java for that matter, are probably null reference exceptions. If we had had a stronger type system that would allow you to say that ‘this parameter may never be null, and you compiler please check that at every call, by doing static analysis of the code’. Then we could have stamped out classes of bugs.
Instead of being able to send null DateTime, the WS dictates that the client must send something. Therefore, the client sends Epoch date thus the code must handle this properly.
What about dealing with Database when NULL actually means something (in certain situation, we do want to put NULL instead of some random value/junk).
IIUI Anders isn't suggesting that null isn't necessary, but that you should be able to say in code: this type does not allow nulls (or this type does allow nulls). We have this with value types, int or int?, but not with reference types. Adding this in at the start would have removed many chances for bugs.
E.g. I frequently need to check if params are null and then throw an exception at run time. If I could declare a param as requiring a value then the error could usually be caught at compile time.
I do not see the need to constantly evolve a language and/or push for rewrites from scratch. Java as a language succeeded precisely because it was relatively simple, yet powerful enough. For example, I believe that adding generics was a mistake, especially in a crippled form (due to compatibility with previous language versions).
I favor switching the language/environment once you feel that you need more expressing power. E.g. use Scala, Jython or Clojure if you need JVM, or perhaps Erlang for server-side.
* Lack of first class functions
* 'null' type
C# has fixed both of those.