and to by extension, .NET.  This InfoWorld article looks at Java popularity and usage and some of the challenges it is facing moving forward.  Some truly interesting nuggets..

"Info-Tech sampled 1,900 companies, most of which are midmarket companies with less than $1 billion in annual revenues. The study found that 12 percent of enterprises focus exclusively on .Net as compared to 3 percent focused just on Java. Also, 49 percent center primarily on .Net, compared to 20 percent for Java. " --- That meshes with what I typically see.

"Still, the survey did find that the popularity of .Net decreases very gradually as the size of enterprises increases. But Goodall cautioned that in such companies, .Net's popularity decline did not come from an increase in Java usage, but instead from a preference for other development platforms in heterogeneous environments." --- I am not sure I would entirely agree with they "why" proposed for the differences between Java and .NET usage as enterprise size increases.  My experience has been that most larger companies lean more towards Java mostly because that was the first place Java penetrated and became a standard.  Big companies with big application needs five or six years ago predominately turned to Java.  They now have a large Java installed base.  It is not easy to try and change the course of that ship.  I rarely find large Java installs that have Java apps being deployed on multiple platforms - it happens, but it's rare. 

"Hansson agrees that .Net is also threatened by new frameworks, but he noted that .Net nevertheless seems to be taking away mindshare from Java in shops predisposed to use Microsoft technology." --- True to the point that both Java and .NET need to adapt to the changing development landscape.  Having said that, I think .NET is in some ways better positioned to adapt to these changes.  Things like the DLR, Project Astoria, etc., show we can adapt the platform quickly to try and incorporate the new things that are coming from other languages and platforms.  There seems to be some similar types of efforts on the Java side, but they seem more ad hoc and "bolt on" than something really happening closer to the heart of the platform.  Just my impression and I haven't spent a lot of time looking at it in detail so maybe I am wrong.

All in all, the article is an interesting data point.