If people care so much, why isn’t there an “OpenJava” language built around OpenJDK? It won’t be a “real” Java (because of the compatability testing suite thing, IIRC) but if Oracle is letting it fester, at least it’d be moving.
I imagine something like that will happen if/when it becomes clear that these rumors are valid. Right now I think the question of Oracle’s commitment to Java is still somewhat open.
That said, this definitely worries me, as all of our products are built on a JVM stack, and a world where the JVM disappears or falls far behind the times, would be a bad place for us. Fortunately since OpenJDK is open source the worst case is probably not as dire as it might be, but this definitely has me worried.
OTOH, as big as Java is, you almost have to assume that somebody would step in and commit serious resources to keeping it alive if Oracle walks away.
I hope so anyway… I like a number of other platforms and languages, but there are none that I relish porting our entire product line to.
OTOH, as big as Java is, you almost have to assume that somebody would step in and commit serious resources to keeping it alive if Oracle walks away.
The problem is exactly that – its size. There are only a handful or organizations in a place to take something of that size and momentum and not have it rip them apart… and a lot of them have their own java alternatives. I suspect the result would be a lot of half-baked much smaller forks. “Java /w Foo” “Java sans Bar” run by small groups… this might be good or bad, but it would obviously have a major impact on enterprise faith.
That’s definitely a possibility. But I suspect an IBM or Red Hat or somebody of that ilk (or some combination thereof) would keep a “mainstream” Java around for the foreseeable future. But, I could be wrong. It’s definitely a concerning situation, no doubt.
OTOH, another recent article suggests that Oracle is actually doubling down on Java support and is pushing for expanded support for Java on mobile devices, including even iOS and Android support. If that is true, I doubt Java is going anywhere in the near future.
As I was reading this article, that’s exactly what I thought: if these rumours pan out, people will fork Java. And I’m not sure that would be a bad thing.
There used to be tons of Java implementations around (Apache Harmony, Cacao, JamVM, Jikes, …) and at least two alternative standard library implementations (GNU Classpath and Apache Harmony).
While some of them are still alive and well and target niches not covered by OpenJDK and HotSpot, a lot of them got the wind taken out of their sails because Sun moved Java towards an open source platform in earnest.
I’m pretty sure that if Java gets neglected further by Oracle, some fork(s) will be developed. They won’t be named Java, but that doesn’t matter much anymore anyway - look at Android.
That’s a fair point, and while Oracle might be uninterested in developing Java or suing open source projects, that could change if there is some perceived value in either of those activities.
Oracles track record would also deter adoption of any such fork by larger corporations.
However, Oracles defeat in the Oracle-vs-Google suit has removed most of the ammunition for any potential future case.
Defeat? They more or less lost the patent case and the six-lines-of-copied-source case, but they managed to get an insane judgment on “APIs are copyrightable”. That’s pretty much the worst outcome for software developers in general.
Oh yeah, I forgot about that. That’s insane, you’re right, and it means that you can’t rely on any non-copyleft licensed APIs, especially not those from Oracle.
However, OpenJDK and it’s libraries are licensed under a copyleft license, and as such don’t qualify for copyright lawsuits.
Isn’t a lot of that around the TCK and the rights to actually call the result “Java” though? If so, it seems that the industry could fork OpenJDK, rename the ecosystem “Kajulari” and make the project OpenKDK and just move on without Sun/Oracle. Yes, technically it wouldn’t be Java anymore, but if wholesale adoption of the new thing occurred, it would become the de-facto successor to Java.
Which is exactly why Oracle is happy about the “APIs are copyrighted”. Your clone uses the same API (which of course it would as a Java implementation)… copyright violation, new stick for Oracle.
But as long as you comply with the licensing terms for OpenJDK (GPL + Classpath Exception, IIRC), then copyright shouldn’t be an issue.
OTOH, doing a cleanroom implementation ala Apache Harmony, would be an issue.. and, if memory serves, I think some Harmony code was involved in this whole Dalvik controversy. :-(
No, the recent ruling, if it gets taken as a precedent for other courts, would hold OpenJDK to be a derivative work because it shares method names and signatures with Oracle’s version.
This might be difficult for Oracle to make use of, because that ruling was issued by a court that usually handles patent cases, not copyright cases. But it is, in theory, the precedent they can now point to.
No, the recent ruling, if it gets taken as a precedent for other courts, would hold OpenJDK to be a derivative work because it shares method names and signatures with Oracle’s version.
Sure, but Oracle are the ones who released OpenJDK under the GPL + Classpath Exception. So, theoretically, as long as you comply with the GPL, Oracle wouldn’t have any (copyright based) grounds for any legal action.
They lost the patent cases because Android bends over backwards to implement things in a different way from OpenJDK. Anything forked off from OpenJDK wouldn’t be able to claim that defense; it would be an open-and-shut case against them.
The issue is that Oracle has shown that even completely separate implementations are not save from harassment and bogus lawsuits, so why should OpenJDK be?
Yes, it’s free/libre/open-source on paper, but Oracle managed to get so many strings attached that it hardly qualifies in practice.
The patent grant that Sun put in place and Oracle must grudgingly honor only protects OpenJDK, not any forks thereof, so the copyleft license is effectively pointless.
This is a ridiculous opportunity for Microsoft, and I say that despite having been around during their world domination attempts. If they can get clr working well across platforms, suddenly that’s a valid system to jump to.
A FreeBSD developer was trying to get Oracle to certify Java on FreeBSD. I wonder if that’s why it stalled? I don’t have all the details, but that would be annoying if this is part of the reason why it hasn’t happened.
If people care so much, why isn’t there an “OpenJava” language built around OpenJDK? It won’t be a “real” Java (because of the compatability testing suite thing, IIRC) but if Oracle is letting it fester, at least it’d be moving.
I imagine something like that will happen if/when it becomes clear that these rumors are valid. Right now I think the question of Oracle’s commitment to Java is still somewhat open.
That said, this definitely worries me, as all of our products are built on a JVM stack, and a world where the JVM disappears or falls far behind the times, would be a bad place for us. Fortunately since OpenJDK is open source the worst case is probably not as dire as it might be, but this definitely has me worried.
OTOH, as big as Java is, you almost have to assume that somebody would step in and commit serious resources to keeping it alive if Oracle walks away.
I hope so anyway… I like a number of other platforms and languages, but there are none that I relish porting our entire product line to.
The problem is exactly that – its size. There are only a handful or organizations in a place to take something of that size and momentum and not have it rip them apart… and a lot of them have their own java alternatives. I suspect the result would be a lot of half-baked much smaller forks. “Java /w Foo” “Java sans Bar” run by small groups… this might be good or bad, but it would obviously have a major impact on enterprise faith.
That’s definitely a possibility. But I suspect an IBM or Red Hat or somebody of that ilk (or some combination thereof) would keep a “mainstream” Java around for the foreseeable future. But, I could be wrong. It’s definitely a concerning situation, no doubt.
OTOH, another recent article suggests that Oracle is actually doubling down on Java support and is pushing for expanded support for Java on mobile devices, including even iOS and Android support. If that is true, I doubt Java is going anywhere in the near future.
A kind of perversion of the original Java promise: write once, run everywhere…
… it’s getting closer … keep running.
(Sorry if that was triggery for anyone! I don’t think it is for many people, but it’s kind of borderline. And sorry for the frivolity, as well. :))
As I was reading this article, that’s exactly what I thought: if these rumours pan out, people will fork Java. And I’m not sure that would be a bad thing.
There used to be tons of Java implementations around (Apache Harmony, Cacao, JamVM, Jikes, …) and at least two alternative standard library implementations (GNU Classpath and Apache Harmony).
While some of them are still alive and well and target niches not covered by OpenJDK and HotSpot, a lot of them got the wind taken out of their sails because Sun moved Java towards an open source platform in earnest.
I’m pretty sure that if Java gets neglected further by Oracle, some fork(s) will be developed. They won’t be named Java, but that doesn’t matter much anymore anyway - look at Android.
And then they will run into exactly the same trouble as Android, except that most efforts don’t have Google’s resources when it comes to litigation.
That’s a fair point, and while Oracle might be uninterested in developing Java or suing open source projects, that could change if there is some perceived value in either of those activities.
Oracles track record would also deter adoption of any such fork by larger corporations.
However, Oracles defeat in the Oracle-vs-Google suit has removed most of the ammunition for any potential future case.
Defeat? They more or less lost the patent case and the six-lines-of-copied-source case, but they managed to get an insane judgment on “APIs are copyrightable”. That’s pretty much the worst outcome for software developers in general.
Oh yeah, I forgot about that. That’s insane, you’re right, and it means that you can’t rely on any non-copyleft licensed APIs, especially not those from Oracle.
However, OpenJDK and it’s libraries are licensed under a copyleft license, and as such don’t qualify for copyright lawsuits.
If you use OpenJDK, which is exactly what Oracle wants, because they can just keep increasing the pain until you give up and buy a license from them.
Isn’t a lot of that around the TCK and the rights to actually call the result “Java” though? If so, it seems that the industry could fork OpenJDK, rename the ecosystem “Kajulari” and make the project OpenKDK and just move on without Sun/Oracle. Yes, technically it wouldn’t be Java anymore, but if wholesale adoption of the new thing occurred, it would become the de-facto successor to Java.
Which is exactly why Oracle is happy about the “APIs are copyrighted”. Your clone uses the same API (which of course it would as a Java implementation)… copyright violation, new stick for Oracle.
But as long as you comply with the licensing terms for OpenJDK (GPL + Classpath Exception, IIRC), then copyright shouldn’t be an issue.
OTOH, doing a cleanroom implementation ala Apache Harmony, would be an issue.. and, if memory serves, I think some Harmony code was involved in this whole Dalvik controversy. :-(
No, the recent ruling, if it gets taken as a precedent for other courts, would hold OpenJDK to be a derivative work because it shares method names and signatures with Oracle’s version.
This might be difficult for Oracle to make use of, because that ruling was issued by a court that usually handles patent cases, not copyright cases. But it is, in theory, the precedent they can now point to.
Sure, but Oracle are the ones who released OpenJDK under the GPL + Classpath Exception. So, theoretically, as long as you comply with the GPL, Oracle wouldn’t have any (copyright based) grounds for any legal action.
Now patents, that might be a different situation…
They lost the patent cases because Android bends over backwards to implement things in a different way from OpenJDK. Anything forked off from OpenJDK wouldn’t be able to claim that defense; it would be an open-and-shut case against them.
That’s already the case with OpenJDK.
The issue is that Oracle has shown that even completely separate implementations are not save from harassment and bogus lawsuits, so why should OpenJDK be?
Yes, it’s free/libre/open-source on paper, but Oracle managed to get so many strings attached that it hardly qualifies in practice.
Could you explain what those strings are?
The patent grant that Sun put in place and Oracle must grudgingly honor only protects OpenJDK, not any forks thereof, so the copyleft license is effectively pointless.
This is a ridiculous opportunity for Microsoft, and I say that despite having been around during their world domination attempts. If they can get clr working well across platforms, suddenly that’s a valid system to jump to.
The .NET team is salivating reading this article…
“Java has no interest to [Oracle] anymore”
Not quite true. Their legal folks will be interested in it now and well into the future.
A FreeBSD developer was trying to get Oracle to certify Java on FreeBSD. I wonder if that’s why it stalled? I don’t have all the details, but that would be annoying if this is part of the reason why it hasn’t happened.