The ASF Resigns From the JCP Executive Committee | Hacker News Reader

The ASF Resigns From the JCP Executive Committee

blogs.apache.org

260 pointsdavidw57 comments

Open articleView on HN

For people who are unfamiliar with all the acronyms:

ASF : Apache Software Foundation

EC : Executive Committee

EE : Enterprise Edition

JCP : Java Community Process

JSR : Java Specification Request

JSPA : Java Specification Participation Agreement

SE : Standard Edition

TCK : Test Compatibility Kit

That helps, but can someone provide a babblefish translation of what this means for those of us who don't live in the Java world? Is it basically...

Dear Java,

We remember when you used to be cool. Now you spend all of your time drinking and hanging out with lawyers. We are leaving and taking the dog and all of the CDs with us.

Love, Apache

Translation :

We are breaking up, except this time, it's not me, it's you

-Kicking ass and taking names

I believe that is basically it. They won't be participating in writing specifications, but I don't think that anyone will be stopping Apache (or anyone else) from implementing any of the JSRs.

If a hypothetical Java Servlet API v4.0 would be released they could still make Tomcat be compliant with it.

On the other hand Tomcat is very widely used in the not-quite-huge-enterprise companies and it is obviously helpful for a spec if all implementers can agree on some common ground before it is written. ECMA probably wouldn't see a point in doing ES5 if Microsoft and Mozilla would not talk to them.

It might be a matter of time before Oracle decided to cut all of the oxygen lines for these open source solutions judging from Ellison's comment about RedHat business models.

What do you think would happen in order for a library to be certified (passed a certain TCK) would require to pay Oracle a huge amount of money?

Oracle can get more money from IBM, RedHat, SpringSource (SS uses a modified Tomcat for their tcServer) and a slew bunch of other companies.

Whether they'll do it or not, I don't know.

Dear Java,

You said you wanted to be friends, but Oracle just wants us as fans. We need more in a relationship.

Love, Apache

Another important one is this:

FOU: field-of-use restrictions (in the TCK license).

http://skife.org/java/jcp/2010/12/07/the-tck-trap.html

http://jcp.org/en/jsr/results?id=5111

Right that is the big one that I have been explaining to people when asked. The secret under the covers is mobile and Oracle's dead set position on that platform. The are doing everything in their power to eliminate all non-Oracle run times from mobile devices. With mobile expected to be the dominate platform in 18 months and it's relative immaturity when compared to other older platforms, there is a huge potential for revenue to anyone who can become the Microsoft of mobile.

Now the conspiracy theorist in me thinks that Oracle is eliminating the VM's on mobile not because they have competency there (Ellison is smart enough know it would be a forgone battle to try to now edge in between IOS and Android) but rather to clear the market for a sympathetic partner that will split the purse. And I think that partner is none other than Apple.

I find the timing of Apple relinquishing the JDK back to Oracle timely and suspicious further both Steve and Larry know that the other has significant competencies is separate, independent, non-overlapping, but complementary markets. If the mobile market it seceded to a closed vertical vendor they are free to chose who become the infrastructure to support the new global mobile network. I think the land has already been divided up and now it is just time to play the battle out.

For people confused about the impact of this:

Having Apache on the EC (Executive Committee) strengthened Java by giving an official voice to the (large) open source Java community. This was useful for Java because Apache often agitated to make sure specifications were licensed under terms that are compatible with open source implementations. Open Source implementations have kept Java competitive with .NET in terms of price, and many specifications have grown out of open source Java projects.

The Eclipse organisation remains on the EC, so Oracle can point to them as a voice for open source. However, Eclipse is different to Apache in that it is primarily a pay-to-play organisation, whilst Apache is a meritocracy.

In terms of specifications themselves (JSRs), Apache will no longer automatically have a representative. Individual experts can still be invited, but Apache's withdrawal (as well as that of people like Doug Lea & Bob Lee) makes it less likely experts will want to serve on a JSR committee.

However, Eclipse is different to Apache in that it is primarily a pay-to-play organisation

This isn't related to the discussion at hand, but can you elaborate on this?

The Apache Foundation board was initially made up of people involved in creating Apache software, and membership is expanded through election by the existing board.

The Eclipse Foundation board is mainly representatives from companies who have paid the substantial "strategic member" fee.

First: Further, the project communities of the ASF, home to Apache Tomcat, Ant, Xerces, Geronimo, Velocity and nearly a 100 mainstay java components have implemented countless JSRs and serve on and contribute to many of the JCPs technical expert groups.

And then:

To that end, our representative has informed the JCP's Program Management Office of our resignation, effective immediately. As such, the ASF is removing all official representatives from any and all JSRs. In addition, we will refuse any renewal of our JCP membership and, of course, our EC position.

Holy crap! That just sounds like Java is really bleeding now. Anyone knows how significant the real impact of leaving representatives on the JSRs is?

Not really. JSRs have never been leading the innovation in Java world: Ant, Maven, Velocity, Spring, Hibernate, Grails, Hadoop, GWT, Android, Struts, Wicket, Eclipse - none of them started as a JSR.

The only JSR that matter is JavaSE and I don't see it having technical problems. JavaME is dying and Spring showed JavaEE how things should be done so they are catching up now.

A counterpoint: On the technical front, I see JavaSE/the JRE falling hopelessly behind C# and the CLR -- fear of Microsoft and distrust of Mono have kept the playing field a bit closer to being level, but I don't see Oracle as an organization that can really take advantage of that.

You are right, and the fear & distrust go both ways. The .NET "community" is almost entirely Microsoft fans and a good number of shills. The shills are rabidly anti-FOSS, the Microsoft employees are about evenly divided between FOSS-hostile and MS-provincial (with a tiny minority pro-FOSS), and the rest are just unaware of anything that Microsoft isn't hyping.

It's very hard to get momentum behind a free software project on .NET, and once you do Microsoft and their shill army kills it by reinvention, heavy promotion of their vaporware, and badmouthing the competition.

It's really hard to work against the culture, and the dominant .NET culture sees Redmond as the One True Source for anything worth using.

fear of Microsoft and distrust of Mono have kept the playing field a bit closer to being level, but I don't see Oracle as an organization that can really take advantage of that

Honestly, as someone who has no interest in Microsoft technologies and distrusts Mono, but who likes the idea of a batteries-included, CTO-friendly, cross-platform, JIT-compiling VM with a huge userbase and oodles of libraries, I am feeling a little lost now. Not that Java is going to die quickly -- I'm still learning Scala and expect that to be a good investment for a few years to come -- but it's starting to look like Java is a platform with a lot less future than I expected.

I'm still learning Scala and expect that to be a good investment for a few years to come

First let me disclaimer that this is totally my perspective on the subject but as a decision maker in a company that is now looking to move on and believes that this is the fatal shot to Java, I wanted to say that for us we feel that with the JVM shenanigans that any technology that runs on the JVM is in danger of Oracle's legal reach. After kicking this one around a lot among our peers we are hopeful that a language that run on the LLVM makes it to the forefront as a replacement.

It will be many years before a new platform can match the libraries available on the JVM. Such a loss :-(

Yep needless to say, I am pretty concerned right now. I think the irony of the whole situation is that much of the community expected this from Mono.

The only JSR that matter is JavaSE and I don't see it having technical problems.

I don't know what you mean by that. Just have a look at what JSRs JDK 7 consists of:

http://openjdk.java.net/projects/jdk7/features/

Now imagine some of Apaches representatives were spec leads on some JSRs. Thats a serious threat to a JSR then. Spec leads are not easy to find. Participators in a JSR can go with important knowledge on the topic, too.

Java SE is almost 100% designed and implemented by Oracle. (The only exception I can think of is Doug Lea's concurrency work.) This is a problem in some sense, but also a solution: the JCP was never really a bottleneck for Java SE so its implosion can't delay Java 7/8.

one has to wonder with Microsoft's new found success in (open) Kinect if it won't translate to something more liberating to the rest of their products.

Presumably the already late Java 7/8 will be further delayed.

Practically, I don't think it really matters at all. But it's a very significant symbolic move. It sends a message about where the Java community is now in relation to Oracle.

Of more concern will be Apache's attitude to new development in Java. Will they continue to use it and to make cool new stuff in Java, or is this a breaking point for them in terms of using the tech?

     The Apache Software Foundation concludes that that JCP is not an open
     specification process - that Java specifications are proprietary technology
     that must be licensed directly from the spec lead under whatever terms the
     spec lead chooses

So much for Java being an open standard.

I always thought it was odd that ASF, one of the early open-source successes, was so heavily Java which has always been semi-open at best. Stuff like this lends more credence to the ideological purity of the GNU guys..

Given the conclusion (that JCP is not an open specification process), I wonder whether they will move away from implementing JCP specs, and towards other technologies.

Huge loss for the JCP either way.

My, perhaps naive, question is: will there now be a bifurcation of the Java community? Specifically, will there be an open community with its own standards and implementations, and the Oracle community with theirs.

The problem seems to be that you can't simply just go implement something independently and call it, say, "Dalvik", rather than Java, and avoid Oracle's lawyers, due to patents.

Then how difficult would it be to create a new Java that has the features (syntactical and whatever else) that everybody's been clamoring for and calling it a different name? C# basically did that, and it's still around....

C#/.NET definitely infringes Java-related patents. But on the other hand Microsoft has been working on programming languages and related tools ever since the start: they definitely have patents with which they can strike back.

Heck, they are in the same league as IBM when speaking of patents. Only trolls have the balls to go after them.

That's exactly what I said. You can't just give it a different name and do a clean-room implementation, because it still likely violates some patents.

And the patents is where the problems lie. Which is the problem with not being able to pass the TCK - no access to the patent pool.

How does the MS CLR get around this issue? Surely their VM is infringing on some of this pool? Do they now licence from Oracle?

If only it WAS this simple... this hypothetical situation would be awesome!

I am firmly supportive of ASF. I was unhappy with Sun for witholding the TCK, and I am now unhappy with Oracle. Major languages should have multiple high quality implementations. I understand the business issues but something like Java needs to have at least one foot firmly in the 'commons.'

Is Google going to follow suit? They were the other ones to vote against Java SE 7 on the same grounds and seem to share Apache's concerns.

Uninformed question: How does this affect Apache projects and their further development?

Oracle shooting them selfs in the head. The open unrestricted competition in Java world moved that comunity since the begining and everybody got profit from that. The tools were improved and preselected by the user experience only. It was a win win. If your product was not good enough you got the valuable information: that you should to improve yourself. That is the basic foundation of capitalism: don't concentrate on your competitors but on your own products. Now Oracle is trying to block that open market of ideas. They will be the first to get hurt.

“The question isn't who is going to let me; it's who is going to stop me.” (Ayn Rand)

As pro-FOSS CTO with a JIT-compiling VM userbase, the ASF exit from JCP EC makes it CYA time for us with the FOU restrictions in the TCK and JDK 7/8. Agreed?

AMEABFTMA (Aaargh my eyes are bleeding from too many acronyms)

IDK, YMMV.

Java by itself as a commercial entity is not extremely valuable. Not really having any clue how to get an ROI out of Java, Oracle like Sun before them is simply hedging their bets that owning/controlling Java IP (of any stripe) may in the future provide commercial returns or business leverage when a Java-based app or framework succeeds (see: the Android lawsuit, the MSVM lawsuit 10 years ago). I guess it's the only logical approach for a company that otherwise has no clue how to monetize a technology as widespread as Java.

Seriously, what did we actually think Oracle was going to do with Java? Get all open and community-minded and crap?

Oracle (or IBM) is profitable enough that they don't have to monetize Java at all.

But then they wouldn't be Oracle...

So what does this actually mean for Java and ASF?

As commentors have mentioned, you can't simply roll-your-own-Java without being at risk of a legal battle.

Zooming out. What are the biggest technologies being affected by this, and what are the alternatives available to the creators and developers of those technologies?

Would we end up in a situation where someone like Google would have to revamp their entire Android runtime?

I'm not sure that revamping would help. Android isn't a JVM, it's already it's own, independently designed and built virtual machine, and yet they're still getting sued by Oracle.

I think you need to re-read the lawsuit.

The claims in the lawsuit would apply to any modern virtual machine.

The acquisition of Java by the intellectual monopolist Oracle is good for programming, because Oracle's rent seeking will discourage programmers from using Java. There are far better languages that have more versatile typing operations than mere subclassing.

I don't think this was wise of Oracle. They only seem to view Java as an asset they can make money of. But I don't think it's wise to turn the tide against them for a few cents.

Maybe Oracle's doing the right thing for Java afterall. Assuming Ownership.

Is their blog down?

Yeah:

"So much traffic/interest in the ASF leaving the JCP that blogs.apache.org is wonky. Oh yeah, it's Java. #JCPIsDead"

http://twitter.com/#!/jimjag/status/12939142254034944

I had to do a bit of searching to figure this out, but that twitter quote is from the President of the ASF - interesting.

Hysterical, cheap shot though it is :) Thanks for sharing.

Yeah, here it is in their mailing list archive, though:

http://mail-archives.apache.org/mod_mbox/www-announce/201012...

Edit: the blog post is back up.

I have put the blog post in pastebin http://pastebin.com/gcsjzkKk if you cannot read it

Reply on news.ycombinator.com