I've never quite followed why unrelated actions are thought to be somehow indicative of how Oracle is or is going to manage the JVM. Assuming you were going to be using their implementation anyway, extracurriculars like this Hudson drama seem entirely unrelated -- especially given the continued emphasis on OpenJDK (which has a GPLv2 + classpath license).
JVM languages like Clojure, Scala, JRuby, etc. would seem to be an entirely safe choice AFAICT. I'd be interested in hearing concrete contrary theories.
If the Hudson debacle was an isolated case, I'd agree with you. But Oracle have been alienating a lot of people in the open source community, and this incident isn't going to reassure anyone of Oracle's intentions.
And whilst the OpenJDK is open source, Oracle have made it pretty clear that they are not above using trademarks and patents to enforce their control over the JVM.
I guess my response would be, "So?". The question is whether the JVM is a reasonable, safe, and reliable platform to build upon. While Oracle's actions vis à vis various projects haven't been what many would have preferred, AFAIK they've stated their intentions for OpenJDK pretty clearly -- and the recent folding of IBM and Apple into that project is no small endorsement. As for control of their VM, yeah, I'd expect that (the Google issues are entirely separate from the JCP drama, etc, which is SOP since long before Oracle came around).
FWIW, I say most of this to reassure myself more than anything else, given my investment in the JVM. If there are real substantive reasons to be concerned – essentially, indications that Oracle has likely just lied about their plans for the JVM, or something more than the vague warnings I usually see – I want to be the first to know.
The JVM is safe for precisely as long as it's in Oracle's interests for the current situation to pertain. Oracle aren't capricious per se, but they are the very definition of capitalism, red in tooth and claw, in a way that Sun famously weren't. If they see a way to monetise the platform that as a byproduct happens to close it down as an open venue, I don't think that would stop them for a second.
I don't think the JVM is in any particular danger of disappearing, otherwise I wouldn't be still programming in it :)
However, a technology controlled by a single vendor, even one as large as Oracle, is at greater risk than one that is an open standard with several competing implementations. I'd be happier if Oracle were encouraging alternative JVMs, rather than suing them or withholding testkit licenses.
It's a function of stability: I want to limit the number of variables that can impact my company. If things go south between Oracle and the rest of the open community, it will cause distraction as the dev community switches to the open VM (with no guarantee it is 100% compatible thanks to no compatibility tests) and I switch over to the open VM (with no experience as to its scalability and stability).
There is no doubt that a year later, all would be fine, but in that year, my pain was increased (potentially, obviously). Would I re-write code for this potentiality? Not a chance. Does it factor into my decision for a new architecture? Absolutely.
The short of it is that I don't trust Oracle. I haven't for many years, but I was willing to give them the benefit of the doubt when they purchased Sun in hopes Sun's culture might impact Oracle's. They have proven they are going to behave exactly as I expected them to. My company won't have infrastructure from companies I don't trust.
JVM languages like Clojure, Scala, JRuby, etc. would seem to be an entirely safe choice AFAICT. I'd be interested in hearing concrete contrary theories.