Read the rather startling comments to the article The long-running Sun-Apache dispute at Javaworld.
The background is that Sun twice started the process of submitting Java to become a standard (once through Ecma, once through ISO, IIRC) only to withdraw. Finally it came up with its own Java Community Process (JCP) where a wide variety of Java Specification Requests have been made (equivalent to W3C Recommendations). The JCP is widely seen as a way for Sun to get community involvement and the benefits of openness while retaining an effective Sun veto.
One product of the JCP has been the Java Technology Compatibility Toolkit (TCK): a suite of tests, tools and documentation that determines whether or not a product complies with a particular Java™ technology specification.
It seems that Apache claims they cannot use the TCK to test Apache Harmony (their open source Java platform) because it would impose certain restrictions on use. It seems that passing the TCK is more important than just getting a branding tick: some JSRs are only licensed for systems that pass the TCK.
And here is the startling comment: former Apache chairman and director Greg Stein reports that
The problem? Many of the agreements surrounding the JCK and the JCP, even the operation of the JCP, are covered by Non Disclosure Agreements. This is why Stephen, myself, and others can't provide exact language.
People who think that open processes are not compromised when a single vendor (or a cabal of vendors) can dominate the process, or who think that subversion begins and ends with Microsoft (i.e. if Microsoft is involved it is probably tainted, if Microsoft is not involved it is probably OK), should learn from the inadequate JCP/TCK process.
And legal definitions of "open" need to start include clauses that a standard is not open unless there are no non-disclosure issues with participation: transparency.