I suspect it was neither good nor bad that Apache participated. One good outcome is a ton of AL-licensed core java code implementations, the copyrights of which are not owned by Oracle, and not under their control, easily integrated into most any OSS licensed language.
One bad outcome of the many worthwhile contributions to OpenJDK is that Oracle owns them, they are copyright assigned, and clearly Oracle is not being a good actor in adoption of that code. The whole GPLv2+classpath exception, overloaded with a bevy of patent threats and outright ownership of the code, leaves something to be desired for anyone who champions reuse.
If one were to create the Joe language tomorrow, syntactically different enough from Java and dodging Oracle's patent troves, it would be trivial to adopt all of those AL .jars and extend the language immediately. Not so with the GPLv2 OpenJDK code, forking to borrow the patents is highly suspect, and the code can never be brought up even to GPLv3 and its patent assertions without the owners/copyright holders direct consent.
I sort of view this as a massive failure to the freedom of software perpetrated by Oracle, but no less by the FSF itself, and share my sympathies with all the non-employee contributors to OpenJDK who agreed to copyright assignment. Trusting a foundation such as the FSF with your copyright is one thing, but entrusting it to a for-profit to protect your code for public reuse is a bone headed move.
Of course, all assurances were made by Sun prior to the ASF embarking on Harmony (there was no FoU considerations at that time, that was injected much later in flagrant violation of the JSPA), and prior to their contributing Tomcat to the ASF, that they were moving forwards. Staying with it prior to the Oracle acquisition was questionable, but staying long enough to determine that Sun had polluted Oracle's earlier positions *against Sun* seemed sensible enough. Now that all of this has played out, and the OSS universes of Java, OpenOffice and MySQL all implode, it seems like Apache chose just the right time to exit stage right.
Agreed that .NET is interesting, once all threats of RAND are completely stripped away. MS would be wise to revisit their patent pledges at this time and address their criticisms, it could score them some serious open source credibility in this environment. Especially if they were to contrast themselves to Oracle's JVM ownership. Perhaps the Outercurve Foundation will help to win some of the necessary assurances. Clearly much of the future of computing will exist on portable and multivendor/multi-OSS project VMs.