Sun’s initial response to the Open Letter from Apache. Here is an excerpt:
- Sun is working with as many communities as possible to create an open source implementation of the Java platform under GPL v2 that mainstream open source communities can work with - this includes TCKs.
- Java technology has many stakeholders, and we recognize that we will not be able to please everyone as we move through this process. In some cases, we’ll have to agree to disagree on some points.
- Our current priority is to make the Java platform accessible to the GNU/Linux community as quickly as possible.
- As you’ll note from Apache’s letter, this is a dispute over specific terms, not over Sun providing a TCK.
- We know that the open source process is a journey and we will continue to work with the open source communities and the licensees to determine how Java technology evolves.
A few quick observations that jump out. First, GNU/Linux and GPLv2 are mentioned twice, (possibly) Sun has determined that it doesn’t want to see a JDK under a BSD-style license? This just happens to be the lever that will prevent that?
Second, No one @ Sun has told me this, but from the response it appears that they are digging in “we’ll have to agree to disagree on some points”. We’ll see what the next few weeks bring, but I’m not holding my breath for a response (positive or negative) before JavaOne.
The interesting phrase in that response, which almost everyone commenting so far seems to have overlooked in preference for framing it as an IBM vs. Sun, or ASF vs. FSF, or even BSD vs. GPL issue [1], seems to be "this includes TCKs".
[1] Imagination apparently runs wild when people have trouble understanding what the open letter is actually talking about.
We'll see, I surmise that Sun might budge on this eventually, but that they are more interested in making life easier for GPL'd JDK distros such as yours, Dalibor.
And, once you have a GPL'd TCK, the next question is, does this preclude a BSD-style impementation? Dalibor, you'd be a better one to answer the question, you've been talking to Sun about this recently.
Well, using your method of text analysis, i.e. counting keywords in Sun's response, we arrive at
GPL : once (not twice)
GNU : once (not twice)
Apache : thrice (including the title)
which should tell you that Sun will use the Apache license for the TCK, right? Or maybe that it doesn't want to see a JDK under the GPL, or something equally far-fetched.
I am looking forward to a blog post from someone applying the 'scientific' techniques from the 'bible code' to discover that the whole TCK is in fact enclosed in the response from Sun, if one shuffles and transposes the letters in the right way. ;)
On a more serious note, I'd be surprised if Sun was interested in precluding any implementation of Java, as proprietary implementations extending a BSD-style code base would not qualify for the TCK scholarship, like the ASF does, so their implementors would need to either contribute their code to the ASF, FSF, etc. to avoid the fees, or pay for their own TCK license.
Dalibor, how obtuse? The central issue stands, Sun's currnet priority isn't fixing this, IMO, it is to get past the JavaOne Conf with a focused message. I still suspect that there are going to be licensing issues, will they be solved with JDK 7? yep. I'm assuming this.. But, the question here is whether they are going to bother changing the TCK licensing terms for JDK 5?
Anyway, best of luck with Kaffe.
Well, all I'm saying is that using numerology on Sun's reply to support a GPL vs. BSD argument doesn't really work, beside numerology being of questionable value in the first place.
While it may be tempting to attempt to frame the issue withing the larger context of a GPL vs. BSD fight, as it's an old, familiar mud-wrestling pit, I think such framing doesn't do justice to Geir's more down-to-earth explanation of the issue.
It's not particularly plausible to me that after years of working together with the ASF on BSD-style implementations Sun would have an interest in precluding such implementations. They've heaped praise on Harmony ever since we started it, have featured it regularly at Java One, etc.
The core of the problem, in my opinion, is that just a few JSRs had spec leads that managed to secure the release of both the RI and the TCK under an open source license so far.
Changing that situation won't happen overnight, but seems to me to be more rewarding in the long term than arguing about details of proprietary licenses of old TCKs over and over again for each of the two dozen of GNU Classpath-based free VMs, for example, in particular as long as GNU Classpath is not a full 1.6 implementation yet.
As one of the few guys who founded Harmony, I'd be very happy to hear it passes any official compatibility test suite ASF can get its hands on, l.0, 1.5, 1.6, whatever, but as I am not a member of either organization, I can only speculate about either organizations plans or motivation.
I just don't think that the answers you're looking for can be found in simplistic conspiracy theories a la 'omg!1! teh ASF's hatin' on Sun for using teh GPL' or 'omg!1! teh Sun's hatin' on ASF for usin' teh BSD', which sums up most of the reactions I've read so far to the open letter in places like TSS and JL.