Java 7 JSR Passes While Transparency Concerns Prevail

Discussions

News: Java 7 JSR Passes While Transparency Concerns Prevail

  1. Stephen Colebourne posted "Java SE 7 passes in the zombie JCP" today. In it, he points out that Java 7 (JSR 336) has passed the JCP process (yay!) but says that the process is really broke, and everyone knows it, what with 13 yes votes, one no vote, and one abstain votes - and six of the yes votes said the vote was on technical merit or a desire to move forward and wasn't an approval of licensing.

    The abstention (Werner Kiel) was due to licensing, and so was the "no" vote (from google). Basically, every comment was referencing the lack of transparency on the part of the JCP, or the license terms that prevent Apache Harmony from moving forward. None of them had any issues with Java having a new revision.

    What does this mean? People love Java. Everyone on the JCP thinks it's important to keep Java moving. Thats great, really; it looks like the JCP is in harmony from a technical standpoint.

    But eight votes - a majority in a 15-vote situation - are protesting the terms. Oracle has said they'd move on with Java SE 7 regardless of what the JCP said, so we'd have a SE 7 irregardless, but everyone's looking at the emperors clothes and saying "we might as well go along with it because not going along with it hurts java more than it helps."

    At least, thats the way it look's to me. If the JCP voted no, the political fallout wouldn't prevent Java 7 from being released. Oracle's already said so. But it would cripple the JCP in terms of how people relied on it, which would mean that the only standards body that exists for Java would be trashed by perception instead of just in reality.

    What I mean is: the JCP is already crippled, Oracle's willingness to release SE 7 despite the JCP is proof the JCP has no real power. But people think it does, because it looks like it does, so it still has some weight and merit to it. If the JCP votes against Java SE 7, though, and Oracle releases it anyway, then the JCP loses the impression of power it has, and then we lose what little voice the community has, in the community process.

    It's a good day that SE 7 passed. It's a bad day in the way it passed.

  2. So, first of all Richard, irregardless isn't a word. (Inside joke/reference)

    >>the vote was on technical merit

    Isn't that what the vote should be about? The JCP on Java 7 is designed to get agreement on what should be in Java 7. If someone doesn't like the color of the shirt Larry Ellison is wearing on Tuesday, they shouldn't sit there and say "we love this JCP, but we want to vote no because of something we don't like that isn't part of this JCP."

    Watching Apache vote 'no' on every JCP for the last three years just looked petulant and petty. Let's move technologies forward with the JCP process, and stop using them as a forum to air the dirty laundry about things we don't like about Oracle. Surely there's a better venue for dealing with this issue of licensing. The 'I'm voting no' temper tantrums we've seen over the years by JCP members certainly aren't working, and to be honest, it's becomming boring, irregardless of how you feel about civil disobedience.

  3. I agree with the problems of matching technology and politics. Quite frankly I also personally dislike the culture of entitlement at Apache when even small open source vendors like Caucho have to pay licensing fees for the TCK and Apache gets most TCKs for free (including the Java EE TCK). If licensing for Apache is an issue (and I can agree that it can be from a certain point of view and perdict that the issue will be fixed after the Google/Oracle lawsuit is decided upon), they should deal with it by building consensus in the executive commitee and properly putting pressure on Sun/Oracle (which they so far have failed to do) instead of using misguided strong arm tactics on technical decisions.

    As to transparency, it has largely been a non-issue in many Java EE specifications like JSF, JPA and CDI. There are reforms currently underway to effectively make transparency a non-issue for all JSRs so I really don't understand the point of continuing to whine about it.

  4. Isn't that what the vote should be about?

    A vote should be about the spec and its implementability. That includes license, since if someone imposes on me what license I can use a particular technology with then I have the right to object, or simply ignore what that JCP come up with.

    Watching Apache vote 'no' on every JCP for the last three years just looked petulant and petty. Let's move technologies forward with the JCP process, and stop using them as a forum to air the dirty laundry about things we don't like about Oracle.

    What has this got to do with things people dislike about Oracle? you've just said right there ... last three years ... so during SUNs time also. Maybe you couldn't give a fuss about standards and specifications etc and what terms they come under, but a large number of people do care. Way off boring. More boring is the response of the powers that be in charge of this JCP.

  5. There's a vital points I'd like to add here, which will hopefully alleviate some of the doom and gloom around the JCP.

    (I'll use "I think" when I'm representing myself as an individual expressing an opinion)

    * There are now two JUG members on the EC that can directly pass on concerns, comments, ideas etc from the Java developer community, I think the balance of the JCP has changed for the better :-).

    * Oracle and the other existing JCP EC members are very aware that change needs to happen (see JSR 348 point below).

    * I think that some of the comments about openness and transparency were made in part because the JCP EC members knew that there is going to be reform in this area (http://www.jcp.org/en/jsr/detail?id=348). I know that we (the LJC) would strongly prefer that the openness and transparency ideals behind JSR 348 is 'back ported' (as it were) to existing JSRs like 336.

    * JSR 348 is the first (of a series of JSRs, not sure how many) that are going to tackle openness, transparency and other issues to do with the JCP and the JSR process, see http://www.jcp.org/en/jsr/detail?id=348 for details on this.  JSR 348 is a positive start and we very much look forward to working with Oracle and the other EC members on this.

    * I'll note that it was Oracle who raised this JSR! A positive and welcome move.

    * JSR 348 is being conducted in the open (see http://blogs.oracle.com/java/entry/jcp_next_update).

    There's still much work to be done, but we believe that the JCP is very much worth putting in the effort for.

    Cheers,

    Martijn

  6. Thanks for posting this and your genuine efforts towards positive change. It's tiresome to see all the blanket negavity centered on the odd-bird of Harmony licensing that fails to acknowldge all that is right about the JCP.