Every language that catches on in the market builds up "cruft" over time. Popularity has a down-side as features are tacked on to keep up with the e-joneses.
But a bigger problem is that Oracle is hell-bent on finding ways to sue profits out of ANY Java user, per (ongoing) API fiasco. Who wants to risk being a target of lawsuit greed? True, Microsoft could start doing the same with Dot-Net languages and API's, but so far they are (mostly) behaved. I'll side with the sleeping sharks over the circling sharks.
Either it uses the same API interfaces as Oracle's version, in which case it's a lawsuit risk, OR it has different API interfaces, making actual code incompatible with that written for the Oracle version: a de-facto fork.
> For a bad language it sure is taking a long time to die
Greenfield projects are not abundant within any organization, thus developers invest their lives extending the infrastucture. Consequently, older technologies tend to stay, particularly when they are proven and actually perform well.
And adding to that, just because some tech is new it doesn't mean it's any good. History is packed with the dead husks of many latest and greatest technologies that died off.
