>I think .Net is poised to capture much of Java's current market.

Anyone leaving Java because of Oracle is not going to .Net.

> The C# developers I know really do seem to enjoy the language and associated tooling, and it's very well established in the enterprise.

Doesn't matter. If you are leaving Java because of Oracle, then you don't care about the tooling. Moving to .Net will just move you to another vendor those guys have spendt the last years moving away from.

Sadly, you can't say "anyone" here. I'm quite certain some people will find Microsoft more free than Oracle and, given the surface-level similarity between the two languages, will chose C# before they move on to a different language.

> I'm quite certain some people will find Microsoft more free than Oracle

After all, like professor Wirth once said, P. T. Barnum was right. It's not reasonable to bet your company on the premise either Microsoft or Oracle will have a sudden change of mind and start playing by the rules.

It's not reasonable to bet a dollar on that ludicrous idea.

>>I think .Net is poised to capture much of Java's current market.

> Anyone leaving Java because of Oracle is not going to .Net.

I agree that folks who abandon Java because of Oracle's presence are unlikely to embrace Microsoft, but that doesn't mean that .Net won't continue to erode Java's market share.

Specifically, consider the many shops that employ both Java and .Net. In those environments, I could see .Net's adoption accelerating due to the perception of strife within the Java community: James Gosling and Tim Bray both leaving Oracle, Jonathan Schwartz pitching litigation over Java patents as a revenue stream, the subsequent lawsuit against Google, the OpenSolaris and OpenOffice.org votes of no confidence and subsequent forks, Apache attempting to throw a wrench into the JCP... that's not the picture of a healthy ecosystem.

So it's not that the enterprises will see Oracle and flee, but that they'll see the turmoil and litigation, reason that C# and Java are similar enough, and opt for the more peaceful of the two, at least for new development.

While I love JavaScript, I shudder at the though of using it as a systems integration language. For the verry reason that it shines as a UI development language (flexibility), it makes me cringe to think of using it for system integration work where items like guaranteed delivery and transactions are far more important that flexibility. As well, I just don't see .NET doing it, it has very little adoption amount the unix camps and the next language will most assuredly will come with the blessing of that camp just as Java did. I mean it was the unix guys that adopted it because they could offer there server offerings on their platform of choice while not ignoring the Microsoft juggernaut.

As for LLVM, I have had the same thought and I think you are on to something, I keep thinking whatever it is, it will crop up on LLVM.

More than that, for typical distributed "enterprise-y" type stuff (a crap-ton of services that call each other) MS's stuff is generally superior. Creating a typical WSDL based service on Java using Spring requires a lot more work than with .Net, and all of the nice little features (like tools for manually calling a web-service API) are built-in.

Unless you are a big enough company to be able to roll your own infrastructure tools to make use of linux's advantages or have very significant scale requirements the cost difference is unlikely to be huge (much of the cost will be operations labor anyway) and the service and support backing of MS is a significant boon in the enterprise world.

LLVM and Javascript are contenders but they aren't platforms yet, although I think they will see a ton of use in web apps and web services, especially outside the enterprise.

Bullshit. I work at a .NET shop who had to interoperate with another company developing WSDL based webservice in Java. They could do a whole bunch of stuff that we weren't able to do because Visual Studio didn't support all the stuff that they could do with their open tools (case in point: field level encryption - supported in Visual Studio Pro 2005 but none of later versions of VS that we use)

I'll freely admit I have limited and dated experience of .NET use of WSDLs, but I remember it making me almost cry with frustration. Really, really grim.

I pretty much agree.

Java is under a two-pronged attack. LLVM is quickly becoming the preferred target platform. The harder Oracle is to work with the harder it will be for new languages to target the JVM.

Also Java's niche is getting squeezed from both ends. Python, Ruby, and Javascript can do more of what Java does with less hassle. Languages like Go and Rust are bridging the divide from C/C++ to Java also. Both languages have a nice mix of good performance, predictable performance, high-level semantics, and concurrency.