Open source Java?

Bruce writes "Newsforge is reporting that Java 2 Standard Edition, may soon be set free of Sun Microsystems' notoriously complicated licensing. A group of 12 Apache developers have put together a proposal called Harmony. The proposal appeared as a simple project call last Friday on an Apache incubator mailing list. It would make this new, built-from-the-ground-up version of Java available under the Apache 2.0 free software license. And it's causing quite a stir in the Java community, especially since respected Sun frontmen Tim Bray, Simon Phipps, and Graham Hamilton have given the project their blessing. As yet there has been no reaction from Dr. Java, James Gosling himself, who is in Brazil talking to developers. In a FAQ on the Apache site, Harmony project leader Geir Magnusson Jr. wrote: 'We believe that there is broad community interest in coming together to create and use an open source, compatible implementation of J2SE 5, the latest version of the Java 2 Standard Edition specification. While the Java Community Process has allowed open source implementations of JSRs for a few years now, Java 5 is the first of the J2SE specs that we are able to do due to licensing reasons.'"

Comments

Ugh. Dupe.

[Anonymous Coward]

http://developers.slashdot.org/article.pl?sid=05/0 5/07/1254211&tid=108&tid=8

Re:Ugh. Dupe.

[/ASCII]

Nah, not really. A lot has happened arounf the Harmony project in the last week, and the articles linked in the story gives a decent summary.

I was under the impression...

[Anonymous Coward]

That there was an open source java project already, BlackDown. I fail to see what the big deal is here really, and arn't the Java standards open anyways and anybody could just write up an open JVM/Compiler? I mean nothing is stopping anybody on /. (or in the world) from writing say their own C Compiler, or Lisp compiler, or their own Virtual machine software, once you know the hardware of the target system it shouldn't be complicated to get a "working" emulation/compiler. Now for performance you would probably be better off using an existing solution, but thats why we have Sun's Compiler.

Re:I was under the impression...

[SharpFang]

Blackdown, Kaffe, GCJ, and quite a few similar "branches", all getting somewhere 60% down the way and stopping there. Somehow I don't quite believe the new project will get anywhere near "usable" as well.

Re:I was under the impression...

[k98sven]

Blackdown, Kaffe, GCJ, and quite a few similar "branches", all getting somewhere 60% down the way and stopping there. Somehow I don't quite believe the new project will get anywhere near "usable" as well.

Ok, first Blackdown is 100%. It's not an open source VM. It's a port of Sun's.

Kaffe and GCJ haven't stopped anywhere. Both are using the same class library (GNU Classpath).

Does this look like 'stopping'?

Re:I was under the impression...

[Rylz]

Yes, anybody could start their own compiler, but if they have the backing of a well-known organization with a history for creating good software, I do not see why it would not be a newsworthy event.

Re:I was under the impression...

[bth]

The Jikes Research Virtual Machine is an open testbed to prototype virtual machine technologies and experiment with a large variety of design alternatives. The system is licensed under the CPL. The initial Jikes RVM infrastructure was independently developed as part of the Jalapeño research project at the IBM Research. In 2001, IBM donated the software to the community.

Jikes RVM runs on IA-32 and PowerPC running Linux, AIX, and OS X. The Jikes RVM is implemented in Java and is self-hosted.

There are additional details on freshmeat.net.

Re:I was under the impression...

[Wesley Felter]

Blackdown was never open source.

Even though "anyone can write a Java VM", so far there isn't a complete open source Java VM. If Apache wants to build one I'm not going to complain.

Re:I was under the impression...

[Anonymous Coward]

Yeah, it doesn't work on windows, but who runs THAT anymore.

Re:I was under the impression...

[sbrown123]

Yep. Blackdown was more of a "cool name" for a branch project where a bunch of Sun programmers ported the Sun JVM to Linux. They were even branded, humorously, hackers by Sun. Kinda silly in my opinion.

http://www.blackdown.org/

Re:I was under the impression...

[drew]

blackdown is a community project, but it is not open source, and never has been. blackdown is the original port of the sun jdk source code to linux (and *BSD?), and the source code and binaries are licenced under exactly the same terms as sun's jdk, (which, iirc, essentially *is* blackdown.) in short, binaries are not freely redistributable, and you are severely limited in what you can do with the source code.

Dupes Ahoy!

[jalefkowit]

I liked this story better when it was posted a week ago.

C'mon, "editors". This has to be getting embarrassing. Right?

Re:Dupes Ahoy!

[Richard_at_work]

At least he hasnt deleted this one, my comment history has a number of comments in it that lead nowhere because Zonk deletes stories and comments. Talk about breaking the referential integrety of slashdot :(

Re:Dupes Ahoy!

[geminidomino]

Talk about breaking the referential integrety[sic] of slashdot

You're joking, right?

Re:Dupes Ahoy!

[Anonymous Coward]

Talk about breaking the referential integrety of slashdot :(

Not a problem, slashdot uses mysql!

Re:Dupes Ahoy!

[Anonymous Coward]

Talk about breaking the referential integrety of slashdot :(

Not a problem, slashdot uses mysql!

Bwahahaha, I needed that laugh

gcj and the new license wars

[jgarzik]

gcj is getting pretty close. It sports a full virtual machine, and implements large swaths of awt and swing.

Why start from scratch? It this simply because the Apache folks don't like the GPL?

Re:gcj and the new license wars

[Gr8Apes]

I would guess the answer to that is "yes".

GPL and Apache licenses aren't quite compatible.

Re:gcj and the new license wars

[Richard_at_work]

GPL and Apache licenses aren't quite compatible.

Since the GPL doesnt allow for distribution of code under any other license, then its not compatable with any other license. Other licenses are compatable with it, but its most certainly not a two way thing.

Sometimes different ideologies foster competition, just as Firefox has forced MS to reopen development on IE 7, the GPL license forces people with more broader ideologies to create competitors to GPLed projects.

Re:gcj and the new license wars

[SharpFang]

So you believe this project will make Sun finally to set THE Java Free?
I'm not holding my breath...

Actually, code can be double-licensed under GPL and another license, but you need some conditions for that (i.e. you can't integrate any GPL code into your main tree, you must be the original author of the code (not basing it on any prior GPL code), or backport modifications to your GNU code by others back into the non-GNU version.)

Re:gcj and the new license wars

[Richard_at_work]

So you believe this project will make Sun finally to set THE Java Free? I'm not holding my breath...

I dont believe anything, Im not even a Java developer :) Im reading and commenting on this story because Im waiting for some ObjC code to compile atm, no other reason!

Actually, code can be double-licensed under GPL and another license, but you need some conditions for that (i.e. you can't integrate any GPL code into your main tree, you must be the original author of the code (not basing it on any prior GPL code), or backport modifications to your GNU code by others back into the non-GNU version.)

That doesnt mean the GPL is compatable with any other license, it means that you as copyright holder have the power to license your code in anyway you see fit within boundries (like is your work a derivative etc).

Re:gcj and the new license wars

[k98sven]

Why start from scratch? It this simply because the Apache folks don't like the GPL?

Actually it hasn't been decided if they will start from scratch yet. They might adopt an existing VM. They might adopt the GNU Classpath class library.

The discussions on checking up the inevitable licensing issues are already underway.

Re:gcj and the new license wars

[Hard_Code]

To the contrary, according to recent discussions, GPL and Apache 2.0 are VERY compatible, all except some minor subtle and nuanced corner-case patent issues which the lawyers haven't yet ironed out.

Re:gcj and the new license wars

[Richard_at_work]

You cant redistribute GPL code under the Apache2 license. As my post says, the compatability is one way only, yet only ever is any fuss made about GPL compatability when someone wants to include code under a the GPL.

Re:gcj and the new license wars

[timeOday]

Since the GPL doesnt allow for distribution of code under any other license, then its not compatable with any other license.

What do you mean by that? See Trolltech for examples of software offered under both the GPL and a commercial license, your choice. It makes perfect sense - you can take and not give back if you want, but then you have to pay up.

Re:gcj and the new license wars

[pnatural]

Since the GPL doesnt allow for distribution of code under any other license

Don't you mean re-distribution?

Someone correct me if I'm wrong, but I thought that the GPL only covered re-distribution, and also could not prohibit authors from distributing under multiple licenses. Put simply, as the author of software, I choose the number and type of each license for every release of code.

Re:gcj and the new license wars

[Richard_at_work]

You seem to be unable to differentiate between dual licensing, which has nothing to do with the GPL, and license compatability, which does. Everyone talks a lot about whether a license is compatable with the GPL, but everyone also glosses over the fact that this compatability is one way only - a by product of the way the GPL is written is that it is designed to leech from other opensource projects without the 'paying up' as you put it. My point is essentially that it offers no advantage to the Apache development team to expend any effort in ensuring that their license is at all comptabale with the GPL because they will not foster anything from it.

Re:gcj and the new license wars

[Richard_at_work]

Apologies, yes I did mean redistribution of a derivative work, although I thought that would have been apparent since I was replying to a post that inquired about the compatability of licenses, implying redistribution.

Re:gcj and the new license wars

[stoborrobots]

Since the GPL doesnt allow for distribution of code under any other license

Don't you mean re-distribution?

Someone correct me if I'm wrong, but I thought that the GPL only covered re-distribution, and also could not prohibit authors from distributing under multiple licenses. Put simply, as the author of software, I choose the number and type of each license for every release of code.

You're right in a sense, but the OP is technically more accurate... If you acquire code under the GPL, you can only re-distribute it under the GPL. However, if you write a derivative work based on a work you acquire under the GPL, then the original distribution of that must also be under the GPL.

Re:gcj and the new license wars

[ignorant_coward]

The FSF website says you are wrong. But I don't care. GPL compatibility is all hype and marketing from the FSF camp--it's like any sleezy branding campaign.

Re:gcj and the new license wars

[Anonymous Coward]

Since the GPL doesnt allow for distribution of code under any other license, then its not compatable with any other license. Other licenses are compatable with it, but its most certainly not a two way thing.

And the Apache Software License, version 2.0 is not one of them. Thus, an aggregation of GPL and ASL software can not be distributed under the terms of either license. Thus, they are completely incompatible.

Re:gcj and the new license wars

[killjoe]

People who write code under the BSD or similar "liberal" licenses do so because they WANT everybody to take their code. More often then not some commercial company takes their code and contributes nothing back but sometimes the code will find it's way into GPLed code too.

The people who write this code under the Apache license are in fact inviting all software companies to take their code and do whatever they want with it. If some code based on the GPL also adopts the code at least they will be able to see the end result even if they can't redistribute it or fold it back. If HP or IBM takes it of course they won't even be allowed to look at it.

It seems like a waste though, there is already kafee and classpath and now these guys are going to start a brand new project from scratch. My guess is that they'll try to get IBM or HP to release their code.

Re:gcj and the new license wars

[Dwonis]

the compatability is one way only, yet only ever is any fuss made about GPL compatability when someone wants to include code under a the GPL.

GPL compatibility is important to people because the vast majority of FLOSS projects are licenced under the GPL. So, if you are releasing a FLOSS program that you want to be useful to the majority of FLOSS developers, you probably care about compatibility with the GPL, whereas you probably don't care about compatibility with the QPL, XFree86 1.1, or Python 1.6b1 licenses.

Re:gcj and the new license wars

[mdfst13]

"I thought that [the intent] would have been apparent since I was replying to a post that ..., implying ..."

That is not a good approach. On slashdot, you should never assume that anyone has read your parent. It is quite possible for you to get modded up and the parent not (as happened in this case). You should always view your post as existing in a vacuum. Any context that is necessary needs to be included in the post.

Not that I always do that properly either. It's not as easy as it sounds. My point is that one's goal should be to write posts that are self contained. Those who don't will find themselves having to come back and clarify their posts. Just like you did here :)

Re:gcj and the new license wars

[Dwonis]

My point is essentially that it offers no advantage to the Apache development team to expend any effort in ensuring that their license is at all comptabale with the GPL because they will not foster anything from it.

Right, because any idiot can clearly see that Harmony has nothing to gain from gcj or GNU Classpath.

Re:gcj and the new license wars

[Krach42]

http://www.gnu.org/philosophy/license-list.html

"Whether it is compatible with the GNU GPL. (This means you can combine a module which was released under that license with a GPL-covered module to make one larger program.)"

You can combine two projects under different licenses into a single product, and have parts under the GPL, and other parts not under the GPL.

Now, I don't know who you're talking to, but I think I'll trust the people, who actually wrote the GPL.

Re:gcj and the new license wars

[Anonymous Coward]

People who write code under the BSD or similar "liberal" licenses do so because they WANT everybody to take their code. More often then not some commercial company takes their code and contributes nothing back but sometimes the code will find it's way into GPLed code too.

You fail to consider a possibility regarding the BSD license: let's say you're a developer. You have a job at a company. If you develop code under a BSD license, you can enrich your company with the extra edge you provide, mingling your code with whatever proprietary extras or extension they want, all the while keeping your code free as in freedom.

The bottom line is: you get to keep your job; your company develops an edge over the competition; and the community still has access to the code. It's fucking great and it's Real World stuff (TM).

Re:gcj and the new license wars

[Anonymous Coward]

And any idiot would realize that it couldn't benefit from these projects anyways, cause they are GPL'd. GPL'd code ABAOLUTELY CANNOT be re-released under any other license unless you are the copyright holder. He is talking about putting Apache code into a GPL'd project, which would be allowed if the code was compatable. Please read the licenses before making a dumb comment.

Re:gcj and the new license wars

[HalWasRight]

Last time I tried gcj, back around 3.2, is was slow as mollases compared to hot spot, which really surprised me since it's native code. Is the performance any better now?

Re:gcj and the new license wars

[killjoe]

if you worked for a company why in heavens would you want your code open sourced? You can just take all the BSD code you want mingle it with your code, provide your company with an edge and keep your job.

Just doesn't make any sense. Every single coder who released their code under the BSD licence just donated hundreds if not thousands of man hours of their time to your company. Take it and run.

Re:gcj and the new license wars

[Richard_at_work]

That only comes into play when you package two individual projects together - in that case the GPL doesnt even come into play for hte other package, compatable or not because its not creating a derivative work of GPLed code. That does not cover redistribution of GPL code in other projects with a different license - completely seperate issue. I shall reiterate - you cannot distribute GPLed code under a different license, the compatability and code flow is one way only.

Re:gcj and the new license wars

[cahiha]

the GPL license forces people with more broader ideologies to create competitors to GPLed projects.

I see nothing "broader" about the BSD or Apache license, and your use of the term "ideology" is misapplied. Code is GPL'ed in order to ensure that all contributors benefit from subsequent contributions. The BSD or Apache license can't ensure that. The fact that the GPL happens to interfere with some commercial or proprietary uses of the software is not ideology, it is incidental.

Re:gcj and the new license wars

[cahiha]

the GPL is written is that it is designed to leech from other opensource projects without the 'paying up' as you put it

The GPL was written when Open Source didn't even exist and there were no BSD and Apache licensed projects at all. The GPL was designed to enable equitable and sustained development of software by a group of people. It does that well. The fact that some BSD and Apache-licensed projects eventually get a GPL slapped on them may be annoying, but it is more a testament to the fact that the GPL is better designed than those licenses than anything else.

Re:gcj and the new license wars

[Darren Winsper]

By taking the code and running, you're essentially creating a fork. If your changes become large enough, it could become difficult to keep up with the main BSD tree, which could simply race off ahead of you.

Re:gcj and the new license wars

[spells]

On slashdot you should never post before reading the grandparent. It is quite possible for the parent to get modded up and the grandparent not (as happened in this case). You should always read the post in context.

My point is that one's goal should be to read posts within the context they are written. Those who don't will find themselves posting poor responses and inane slashdot posting rules. Just like you did here :)

Re:gcj and the new license wars

[killjoe]

So? Look at what apple did. It didn't seem to harm them any.

Re:gcj and the new license wars

[Darren Winsper]

Last I checked, most of the BSD stuff they took is still under a BSD license in OS X.

Re:gcj and the new license wars

[killjoe]

You missed my point. They took a shitload of BSD code, the modified a bunch of it, only gave back a tiny little bit and even then in a form the original coders can't easily use.

It hasn't harmed them a bit.

Re:gcj and the new license wars

[Darren Winsper]

I'm not especially familiar with OS X, but what BSD-licensed utilities have they taken and not released their changes under the BSD license?

Re:gcj and the new license wars

[killjoe]

Virtually all aspects of darwin have been touched by the mac engineers. Just to give you one minor example utilities like tar, rsync etc all were modified to deal with resource forks.

Re:gcj and the new license wars

[Darren Winsper]

Haven't they been released under open source licenses?

Re:gcj and the new license wars

[killjoe]

I'll repeat myself.

Virtually all aspects of freebsd have been touched. Some stuff has been released open source, not all of it in a form the original license holders can use. Most of the changes have not been contributed back.

Re:gcj and the new license wars

[Anonymous Coward]

If your code relies on proprietary extras, your users can't run modified versions, so calling it free is meaningless. If not, the GPL would prevent your competitors from using proprietary extras to get an edge over you.

Re:gcj and the new license wars

[Dwonis]

Open-source projects gain popularity once they become useful. If Project Harmony makes use of gcj or GNU Classpath, it will become useful *sooner*, and thus gain popularity (i.e. more developer attention). If Harmony can't be shipped with gcj or Classpath, it could take much longer.

BSD had benefited from Linux in the same way. I know a number of people who learned about BSD because of the LUG they attended.

It probably feels good to make bold, blanket statements, but it's not terribly useful.

Re:gcj and the new license wars

[Krach42]

*NOT* two individual projects together. It says COMBINE MODULES into one LARGER PROGRAM.

You all are seeing what you want to see. Yes, a derivative work must be released under the GPL, but if you're inserting a module from another license into your GPL code you need not modify that license.

For instance: PearPC is GPL code. But if I took a module under a different license. Let's say under the LGPL (libc) and linked it into my program. Then this wouldn't require libc to change over to the GPL.

Where as, if my program were under a non-compatible license, and I linked in a GPL library (say, librep) then I would be bound to making it GPL.

Conclusion: There exists the ability to combine two licenses into a project, when one of them is GPL. Whether you seem to think that people can't or not.

GPL-Compatible?

[miyako]

If the JVM is licensed under the Apache license, which is incompatible with the GPL, does that mean that you cannot use this with GPLed Java software?

Re:GPL-Compatible?

[m50d]

No, the JVM just runs the code. One would hope the class library will use a license that permits linking any kind of program against it; if not, I'm sure this will die very quickly. No language has succeeded without allowing you to run code under any license at all.

Re:GPL-Compatible?

[Anonymous Coward]

You don't "link" java code to a class library. That pretty much makes your point moot.

Re:GPL-Compatible?

[m50d]

Yeah you do. It's dynamic, but that doesn't matter, it's still a derived work.

Re:GPL-Compatible?

[Anonymous Coward]

According to the fantasy lawbooks of judge RMS. Not according to Sun.

Re:GPL-Compatible?

[Bloater]

Possibly, but you don't distribute the running program, so it is totally irrelevant. Anyway, IMHO, the result of dynamic linking is mere amalgamation - The only effect that doesn't appear to be amalgamation is the precise nature of transient data that is created as a result of the interaction of the two codebases. Since that is not an artistic unity (but rather is a systematic unity), I fail to see how it can be considered derivative art. A computer cannot make art through routine system operation with no particular care put in to creating it. So, eg, a company that provides a service by running a GPL program with a non-GPL class library then sending a JVM core image over the network after the hefty calculations are done would not be distributing a derivative work. Anybody could make the same core image through routine operation - it just might take longer for their weaker CPUs to produce the same image.

Re:GPL-Compatible?

[malraid]

Ok....let's put it this way. I program against Sun's class libraries, using Sun's javac etc... I distribute my propietary (not open source, not GPL, not Apache license) program, but one of my clients decides to run the byte-code that he got from me in an Apache or GPL JVM. How can that be considered derived work, considering that I've never used other JVM apart from Sun's???

Re:GPL-Compatible?

[Brian Blessed]

Q. If the JVM is licensed under the Apache license, which is incompatible with the GPL, does that mean that you cannot use this with GPLed Java software?

A. No, neither license places any requirement on the usage of the software.

- Brian.

Re:GPL-Compatible?

[p3d0]

Actually it's not that simple. Modern JIT compilers do method inlining, and so the jitted code is a derivative work of the class library code, and possibly of some JIT intrinsics.

IANAL, but if you were somehow to redistribute the jitted code including some inlined GPLed code, I think you'd be bound by the GPL to license your own code under the GPL.

However, the usual use of Java is to run the jitted code once, in-place, without distributing it, and then to discard that code when that particular invocation of the JVM terminates. In this usage scenario, I don't think there's any problem with the GPL.

Re:GPL-Compatible?

[m50d]

In that case it's a derived work of Sun's class libraries.

Re:GPL-Compatible?

[m50d]

If their API contains anything substantial enough to be copyrightable, your code could well be a derivative even before it's linked.

Re:GPL-Compatible?

[Bloater]

It doesn't, an API is a description of a language. You cannot copyright a language. It is completely illogical.

Quite a stir?

[Anonymous Coward]

I think this "quite a stir in the community" is wishfull thinking. The Java community at large doesn't care much about an open source Java. People want to or have to write code, not fighting holy OSS wars.

This is a home-made a storm in a teacup. There is already an initiative to create a free Java: GNU GJC. And no one cares about it. The Apache people are just running some propaganda now, but it will be forgotten in a few weeks.

Re:Quite a stir?

[m50d]

The Java community at large consists of people who don't care about open source java precisely because there is no (good) open source java. There's a pretty huge sample bias there.

Re:Quite a stir?

[Anonymous Coward]

Even with one, why use it? Why use a copy if you can have the original? Oh, because it is OSS? Hint: The majority of Java programmers are commercial programmers who give a fucking fart about OSS. They sell their propritary programs and don't fight a holy war.

If the Java community would be interested in an OSS Java, there would be much more support of the existing OSS projects. The only ones asking for an OSS Java and crying out loud about Sun are religious idiots who never have written a line of Java. In fact, these stupid script kiddies would probably not even recognize a Java compiler if bitten by one.

Re:Quite a stir?

[Anonymous Coward]

Agreed. It's painful to see out-of-touch blowhards like ESR begging for Sun to give away Java -- not for the good of the Java world, but for some bogo-strategic "OSS" reasons.

The good news about this Apache group is that they do have an inside track with Sun and might actually be able to convince them to do it.

Re:Quite a stir?

[Bob9113]

The Java community at large doesn't care much about an open source Java.

The vast majority of users of any zero cost technology don't care if it's open source.

While you may be correct about the Java community at large, I think you may be off the mark with respect to senior Java developers. I think a fair number of senior Java developers have, at least once or twice, had language recommendations coldly rebuffed by Sun (and, to Sun's credit, probably have had a few accepted through the JCP). Sun is renowned for sticking their fingers in their ears and chanting, "lalala I cannot hear you lalala" about core language issues. An open source JVM is an opportunity to have our voices heard, or at least to create legally sharable patches for the features we want.

Consider the pre-Tiger dustup over autoboxing; it's a good idea if you want to increase user-friendliness at the cost of strictness and performance. A lot of people didn't want to be forced onto one side of this tradeoff, but Sun's apparent number one priority for 5 was to exceed the perceived user-friendliness of .Net at any cost. Even a command line switch to issue a warning about autoboxing was out of the question.

This is one sort of problem that OSS solves. The vast majority of users of any given technology may not care if it's open source or not, but to the power users it gives a degree of freedom that really matters.

Re:Quite a stir?

[drew]

Even with one, why use it? Why use a copy if you can have the original? Oh, because it is OSS?

I don't care that much if it's oss, but it would be nice to at least have a jdk/jre that is freely redistributable, so that i could install it on freebsd without having to jump through so many hoops. while i prefer open source to proprietary software when possible, i don't have an issue with using proprietary software when it gets the job done better. java in it's current state is a pain to use on most non-linux open source operating systems and that in itself is enough for me to wish for a more open java implementation.

Re:Quite a stir?

[thephotoman]

I care about a free Java, and I do like GNU's GCJ. It's actually quite decent, with all the work that Red Hat's put into it since the OO.o 2 debacle.

I like the idea of being able to have my own site that allows others to download a gzipped tarball of a bunch of free software that includes almost all the dependencies, so that my audience doesn't have to go around hunting down various things (like a JVM) from other places on the Internet.

My only major problem with GCJ that I can think of right now is that I don't know if there's a Windows version, and if there is, how well it works. I may be a GNU user (of course, on Linux), but most of the people I'm trying to introduce to FLOSS are Windows users. That way, I can introduce them to Java programs without any worries about the so-called Java trap.

Re:Quite a stir?

[mikey1134]

I think its important to remember that, while OSS java might not be a big deal to every member of the java community, it is a big deal for java and OSS. A n open source java as a whole. This means that OSS like linux distrbutions could carry JVM's without liscense worries. It also mean the possibility of bring java to platforms Sun does not support on it own. This could negate a lot of the barriers to OSS using java. (Yes I know OSS apps are already being written in java, but those can only move to where java is. If Debian can't carry java, it can't really carry java apps either)

Re:Quite a stir?

[Anonymous Coward]

Auto-boxing changes nothing in terms of performance. If you pass a value type as a parameter to a method that expects an object it will generate code that is precisely the same as if you were wrapping it yourself by hand.

It doesn't even really alter 'strictness' since it basically just promotes value types to their object wrappers, and only where that actually makes sense.

Not to mention that you don't need an open source JVM to enforce your arbitrary language preferences. Indeed, you only need an open source compiler, of which there exists more than one.

Re:Quite a stir?

[radish]

Speaking as a member of the Java community, the reason I don't care about an OS Java (beyond the "oooh neat" factor) is simply that I don't need one. I write Java for work, we need a stable, supported platform to run our apps on. Linux is _just_ getting to where it's useful for us, which is great. Java is there and has been for years. It doesn't cost us anything to run it, we get support, it works. Why again should we replace it?

Re:Quite a stir?

[radish]

I don't think I agree. While it's true that there are people who want Java (the language) to change in some way or another, this is about creating an OS replacement for Java 5. To do that it has to be compatible - you can't change the language because then it ceases to be Java 5. As a professional Java guy responsible for a large installation you're going to have a hard enough time persuading me to drop Sun. You'll have even more trouble persuading me to switch to I-Can't-Believe-It's-Not-Java.

Re:Quite a stir?

[Bob9113]

Auto-boxing changes nothing in terms of performance. If you pass a value type as a parameter to a method that expects an object it will generate code that is precisely the same as if you were wrapping it yourself by hand.

It doesn't even really alter 'strictness' since it basically just promotes value types to their object wrappers, and only where that actually makes sense.

While your statement is true when talking about cases where you intend object conversion, the problem is when you don't want conversion to occur. IBM notes the problem in this article, for the same reason I do. If the compiler uses objects without telling you, it's a problem. Having done a bit of real-time image processing with Java, it is clear that Java is an adequate high performance platform, and that primitives are significantly faster than objects - as long as you can count on them remaining primitive.

Re:Quite a stir?

[Per Bothner]

My only major problem with GCJ that I can think of right now is that I don't know if there's a Windows version, and how well it works.

GCJ does work on Windows. Here is a useful link.

Re:Quite a stir?

[Tim C]

I've been a professional Java developer for 5 years now, and I completely agree. I have never, not once thought "if only Java were open source!". (For comparison, I first started using Linux nearly 8 years ago)

I have no desire or need to change or add features, I have no desire or need to run it on unsupported OSes, and I have no desire or need to distribute it to third parties. I have no desire or, as far as I can see, need for Java to be open sourced.

Your mileage my vary, of course, but for myself, I see no value in it.

Re:Quite a stir?

[Anonymous Coward]

1. That's a developerWorks article, not an IBM article.
2. The compiler never promotes a value type to an object type without you specifically passing a value type to a method that expects an object type of a compatible type signature.
3. The promotion of a value type to an object type will only be relevant in code that already expects an object type. That is, it does absolutely nothing to any code that deals with the value type.
4. That article is about the construction of a completely ridiculous monomorphic collection framework that only exists because the JVM lacks the ability to properly implement parametric polymorphism with specialization. Well no, that's not entirely true. It exists because Java lacked any sort of parametric polymorphism, and only continues to be relevant because it continues to lack explicit specialization and the ability to automatically generate type-safe monomorphic specializations.
5. The article claims that parametric polymorphism in Java hides inefficiency. It certainly doesn't hide it from me.

You cannot accidentally try to stuff a value type into a collection parameterized on an object type, and you cannot accidentally try to stuff a value type into a collection meant for holding object types. I mean you could, but that would be a pretty impressive display of ignorance of your own codebase.

"Gee, I really need Java to tell me that I'm making use of a feature that if I wasn't, this code wouldn't even compile for in the first place. This code needs to be high-performance, but I'm accidentally creating a collection that holds object types and putting my value types within it, and then using it in some compute-intensive area."

Re:Quite a stir?

[Anonymous Coward]

Harmony has relatively little to do with whether or not any particular Java compiler with Bob's modifications conforms to the Java language specification. It's about collecting or constructing the components necessary for providing a J2SE environment. A JVM, an implementation of the class libraries, and yes a conformant compiler.

But Bob there, he can modify the compiler until Hell freezes over and it would mean dick to you.

You just continue compiling your code with your favorite conformant compiler, and it'll run on Harmony's VM.

There are already numerous compilers openly available for the JVM, and I hardly see you ditching the Sun JVM because Bob can compile Nice or Pizza or Jython.

Re:Quite a stir?

[Thunderbear]

One of the major advantages of using Java is the platform independance it provides. This means that you should be able to choose freely the underlying machine and operating system to run your application on.

Currently that means that if you want to run on a PC you need to run either Windows or Linux (or, perhaps, Solarix x86). If you want to run on a Mac you must run OS X - you cannot run just plain Darwin. You cannot run Darwin on a x86. I cannot run Java on my Linux-based router because it uses a MIPS-CPU, and there is no suitable JVM for my needs.

The list is endless.

Harmony might just be what is needed to get the true potential of Java unleashed.

Re:Quite a stir?

[Thunderbear]

What would you do if Sun was bought by somebody, who chose to shut down Java completely? Never provide another release again?

Re:Quite a stir?

[Tim C]

I should probably have mentioned that we do server-side web app development. Even if that happened, we could happily continue using whatever version was current at the time, while planning a migration to whatever replacement technology seemed suitable.

Yes, I've been doing Java a long time, but I was a C/C++ programmer before that, Fortran before that, assembler before, BASIC before that... I'm not so wed to Java that I wouldn't use something else if there was a reason to do so.

Re:Quite a stir?

[ultranova]

Speaking as a member of the Java community, the reason I don't care about an OS Java (beyond the "oooh neat" factor) is simply that I don't need one. I write Java for work, we need a stable, supported platform to run our apps on. Linux is _just_ getting to where it's useful for us, which is great. Java is there and has been for years. It doesn't cost us anything to run it, we get support, it works. Why again should we replace it?

Because currently, without a good FOSS Java implementation, your Java apps live or die at the whim of Sun Microsystems. If they decide to give up on Java, or simply go banckrupt, you're screwed - you get no more support or bugfixes, and ultimately the current JVM will simply not run on tomorrows hardware and OS'es.

On the other hand, if you use a FOSS JVM, you can be sure that it gets updated as long as there's even a single developer interested in Java - you can do it yourself if you have to, or pay someone else to do it for you.

The advantage of free open sourced software is not only freedom, but also security - if you depend on proprietary software, you are at the mercy of the vendor. That open source programs also tends to have higher quality of programming - leading to faster and more stable programs - is a nice bonus too.

Re:Quite a stir?

[Bert64]

And even those of us using linux on non x86 platforms.. The AMD64 version of java has lagged behind the x86 version, and sparc/mips/alpha/ia64 are pretty much screwed..

Re:Quite a stir?

[orasio]

I care about GCJ.
I can develop in SuSE native GCJ binaries for windows deployment, and noone needs to know it's java, other than I, of course.
That allows my team to reuse all the server side software and build windows clients, all with the same language, and the same backend.
That's too good.
Plus, it does work.
We even have SWT binaries working, all that compiled ith good old GCC.

Re:Quite a stir?

[Bob9113]

Hmmm, browbeating tone, steadfast refusal to consider whether there may in fact be a problem here, posting as AC... do you work at Sun perchance? If so, I fear your goal is to win the argument, not to understand the problem (a goal which is well served by your hostile tone), but I'll make one more attempt anyway.

Since you seem focused on attacking the specific form of Dr. Kabutz's example, rather than considering the general problem, I've coded a quick example that is a bit more real-world like.

Autobox.java

Suppose I wrote a posterize function as shown in the above, and placed it as a static method in a color utility class. I then use that method in a section of code that processes 1600x1200 images (yes, I'm sure there are optimizations that could be done to the posterize method above, but let's assume it's close enough, for the sake of argument) (and yes, I know I'm not processing the alpha channel, but work with me here, OK?). All is well, the image processor is working as fast as I need for my case - I'm processing 5 photos per second on a machine just like the one I'm writing from now.

Now suppose some young developer, the very sort of developer Sun is reaching out to with the programmer-friendliness additions in Tiger, comes along and says, "hey, k3wl, we can use autoboxing here and change two of the three parameters in this method signature to Objects without altering any of the invoking code!" And so he changes the signature of the posterize method to the signature you see in my example as posterizeObj. Performance of the method is cut in half, despite zero code changes except the autoboxing.

So, from your ivory tower at Sun (or whatever ivory tower you live in where every developer has a deep understanding of compiler de-optimization), can you tell me where I went wrong? Should I not give apprentices full access to the source code repository? What if he's not an apprentice, just a newly converted SmallTalker who thinks everything should be an object and doesn't understand the performance problem?

Perhaps there should be a comment on the posterize method saying, "Don't change this to autoboxing, it's a high performance method!" But doesn't that beg the question; if, as of JDK 1.5, I have to add more comments to keep other developers from breaking my code without the compiler telling me, is it really more programmer-friendly?

Re:Quite a stir?

[m50d]

Exactly. But the point is if you didn't feel that way, you wouldn't be using java. Which means it's stupid to talk about the community not caring, because the people who do care, like me, are naturally not a part of the community. Who knows how many more people would be using java if it were open source? You certainly can't tell just by asking the current java community.

Re:Quite a stir?

[Anonymous Coward]

You wouldn't expect the language to stop him from bubble-sorting posterColors during each call. Catching dumb design decisions is what code reviews and profiling are for.

Re:Quite a stir?

[Anonymous Coward]

I don't work for Sun Microsystems. I've never worked for Sun Microsystems, and I have no respect for Java as a language and practically none for the design of the JVM.

What that article says that's even remotely interesting (and I say that fairly broadly since that collection framework shouldn't exist at all) doesn't pertain to auto-boxing. And the part that refers to auto-boxing, doesn't support you. Its sole unsupported claim is that it gives a false sense of efficiency.

Well if that's the best piece of support you have from that article, then your entire position already has a strong foundation upon which to stand.

Now onto your example program. Well, what do we see in posterizeObj? Yep, it's an object type. So your argument boils down to "someone that doesn't know what they're doing might change the contract of a class that is key to the performance of our product which may or may not result in a degredation of performance." And what's your rationale for this studious developer? That he thinks "everything should be an object" or has some equally ignorant position?

Well nothing keeps a would-be Smalltalk programmer from enforcing "everything is an object" without auto-boxing, either. It's just not as convenient for them. Big deal? It's a conscious decision to change the interface of a class. Specifically so that a method takes an object type in the place of a value type. It's not slipped onto you. It's not an automatic conversion to inefficiency by javac.

The funniest part of your example is that changing the signature from int to Integer is pointless; instances of type Integer will be demoted to a value-type and the former method will be invoked. Given that instances of type Integer are immutable, and the class cannot be subclassed, there's absolutely no reason to ever explicitly use Integer in an interface exported by a class at all.

It's not less strict and it's not lower performance than if they switched everything by hand. If you don't know that auto-boxed value types are promoted to object types, then you really have no place developing performance-critical software. I'd say that you shouldn't be developing any software all, since contrary to how you like to put it, it's not a "deep understanding" of "compiler de-optimization." It's a basic understanding of the language. What's next, the new for loops are "compiler de-optimization" because someone that doesn't understand the language goes around badly converting every for loop in your program to disasterous effect?

I'm really starting to think that you're retarded.

Re:Quite a stir?

[cahiha]

I think this "quite a stir in the community" is wishfull thinking. The Java community at large doesn't care much about an open source Java. People want to or have to write code, not fighting holy OSS wars.

You are quite right: the people left in the Java community don't give a damn about licenses or even understanding licenses. They blissfully believe that Java is open because Sun tells them so.

You are wrong asserting that the people who care about that sort of thing are fighting a "holy war". By and large, they are people who have been burned by building on proprietary standards before. It's a question of dollars and cents and risk, and Java is too risky.

It's too bad you don't understand yet, but you probably will sooner or later.

Re:Quite a stir?

[cahiha]

OK, so you develop throwaway programs and it doesn't bother you that Sun can do with the platform whatever they like.

But other people develop software that needs to be supported for decades, and for them the current situation is not acceptable.

Re:Quite a stir?

[Anonymous Coward]

>> I have never, not once thought "if only Java were open source!".

I'm on the other side of this; if I'm in a quandary I grab the source code of the jdk to see how something has been implemented. That used to be Ok, but nowadays, thanks to the geniuses at Sun, and, mostly, that horrible 'gang of four', I find myself switching and switching between files; 'Factories' ? 'FactoryImplementations' ? APIs of APIs ? static methods yielding singletons that come out of 'Class.forName()' ? It drives me up the wall, but none of it does it like finding that somewhere, hidden behind all of these layers and layers of obfuscation, the actual implementation goes somewhere to a 'com.sun' class - of which I, of course, don't have to source code.

Anyone sum up...

[SharpFang]

...what are the catches of Apache v2.0 license vs standard Free license like GNU?

Re:Anyone sum up...

[benjamindees]

Seems to be a difference in the way the licenses deal with patent issues.

Re:Anyone sum up...

[m50d]

Patent licenses. It's thought the GPL 3 will be very similar to the apache license.

Re:Anyone sum up...

[punkrockguy318]

The apache 2.0 license has a clause about patents, whereas the GPL does not.

Re:Anyone sum up...

[Vanieter]

Excepted that GPL is copyleft, whereas the Apache License 2.0 is not.

Re:Anyone sum up...

[m50d]

It is copylefted. You have to retain all copyright notices on any redistributed version, including the one saying it's licensed to you under apache license.

Dupe, and why?

[m50d]

Why does this attempt get so much attention? There are plenty of existing attempts at getting a free java, why does apache feel the need to start a new one?

Re:Dupe, and why?

[Anonymous Coward]

Because the Apache marketing department is running overtime. And, because the /. editors have no idea what is going on in the Java community, don't know about existing attempts to create a free Java (not even about the one from the GCC people) and think the Java community is centered around Apache. In reality it is centered around Sun.

Re:Dupe, and why?

[mikaelhg]

We don't just need a Open Source Java, we need a Open Source Java implementation fit for production use.

To produce such a thing, we need a community of competent people committed to that goal. This is what other Open Source Java projects lack.

To get such community going, one needs to communicate in a certain manner. This is what the Harmony people are now doing. The strenght of this project is, to me, that it has both excellent technical competence and competence in community management and in setting and achieving goals in a reliable manner.

This attempt is getting so much attention because senior people who understand that there is more to life than mere technical details pay attention when people who have a track record in producing results, speak.

Re:Dupe, and why?

[mpcooke3]

possibly because there are so many successfull apache jakarta projects.

still doesn't warrant posting it twice, zonk is obviously having a bad day.

Re:Dupe, and why?

[Anonymous Coward]

In other words, the success of Apache represents a triumph of marketing and politics over code.

Re:Dupe, and why?

[Anonymous Coward]

OpenOffice doesn't work with any Java other than Sun's. Sun's version is proprietary, so the zealots don't like using it. Hence the need for a new, Free, Sun-like Java implementation.

Re:Dupe, and why?

[killjoe]

What makes you think the people behind the current open source java projects are incompetent and not committed?

What an insulting thing to say.

Re:Dupe, and why?

[mikaelhg]

Clearly, they are competent, but committed to different goals.

Re:Dupe, and why?

[m50d]

Yeah, but kaffe/classpath are working on that, wouldn't contributing to their projects get the job done quicker than starting a new one?

Dupe...

[Anonymous Coward]

dupe, dupe, dupe, dupe, dupe.....

Helping out current Java Open Source projects?

[Sv-Manowar]

Bringing open source Java runtimes to fruition should be an important step for open source java projects that are currently held back from entering distributions and packages because of this requirement. Also the requirement of Sun Java to use Java 5 on Linux (this situation may have changed..) would be a good thing to challenge.

Having such a fundamental and established organization like the Apache project behind the effort should really aid & help to posture this effort within the wider open-source community.

Re:Helping out current Java Open Source projects?

[Anonymous Coward]

open source java projects that are currently held back from entering distributions

You know, so fucking what? Java is still a huge driver of Linux deployments, probably the #2 reason people run Linux behind Apache HTTPD. There doesn't seem to be any significant reluctance to avoid Java just because it doesn't come bundled in your distro.

Re:Helping out current Java Open Source projects?

[Anonymous Coward]

Jusus. Thanks for that nugget of wisdom!

Re:Helping out current Java Open Source projects?

[LnxAddct]

A non-free JVM hasn't stopped the Fedora devs. They put together GCJ, which is under very active devlopment, but can compile most java applications including apps that use Swing, AWT, and SWT (I believe it supports SWT but never tried) it supports JOGL too so you can code 3D java apps that run natively. In the next release of Fedora, all of the java stuff in Open Office will be compiled with GCJ, same goes for Eclipse and Apache Tomcat. GCJ is an amazing project and being able to run java as native code is even cooler. Look for all this (plus a whole bunch of other cool stuff) in Fedora Core 4 which should be released on June 6th.
Regards,
Steve

Critical thinking...

[PornMaster]

Well, can you run GPLed Java software on the Sun JVM?

Let's get real, folks. Critical thinking isn't that difficult.

Re:Critical thinking...

[Anonymous Coward]

Let's get real, folks. Critical thinking isn't that difficult.

"Ahem..." *points to the name of the site you are on*

Re:Critical thinking...

[malraid]

Do you mean I can also run non-GPLed programs in Linux !? wow.....

Re:Critical thinking...

[Anonymous Coward]

Heretic! Burn him! It's GNU/Linux!

Re:Critical thinking...

[Anonymous Coward]

If you read the posts on this article, you'll see that they are mostly by very junior people with little understanding about the strategic advantages of things such as free software, not to mention bereft of any experience in running Java applications in production environments.

Their lack of critical thinking does not suprise me. Critical thinking requires some experience.

Re:Critical thinking...

[geminidomino]

*Hands malraid a magnum, five boxes of ammunition, a green herb, and the shield key*

You're gonna need these when the Stallman zombies swarm you for that one...

Re:Critical thinking...

[TrappedByMyself]

Critical thinking isn't that difficult.

Neither is trolling, but it's much more fun.

Re:Critical thinking...

[malachid69]

This conversation has actually been on /. many many times.

Although there are some arguments about what the GPL actually says, generally the conclusion is:
1) If the JVM itself is GPL, you could run non-GPL code on it
2) If the core-libraries are GPL, you could NOT run non-GPL code on it (ie: if the Object class alone was GPL you couldn't)
3) If the core-libraries are LGPL, you could run non-gpl code on it (since you are linking to it)
4) most companies are unwilling to use GPL software/libraries becaues they aren't able to use the license restrictions on their own products -- thus, they go for BSD, MIT, or Apache licenses.
5) Most GPL advocates believe the JCP licensing is not open source, though it requires that JSRs can have open-source alternatives written
6) Most JCP advocates believe the JCP licensing is open source, since they don't have to pay to use or implement the JSRs; and because they don't have to use their license for their own products
7) Most JCP and GPL advocates believe that Apache licensing is free/open.
8) Dual-licensing is tricky, and some companies are afraid of using it
9) Most dual-licensing is GPL or Commercial. While dual licensing would allow for GPL or Apache licensing, most GPL advocates would complain that it would not require people to contribute back to the community
10) When people don't like, are afraid of, or can't use existing licenses, they tend to find or foster alternative licensing solutions.

Did I miss anything?

"blessing" doesn't matter

[cahiha]

Sun employees have said many things about Java in the past: how open it is, how anybody can supposedly implement it, etc.

In reality, in 10 years, nobody has managed to create an interoperable, independent implementation of it. All the Java implementations that exist either are highly incompatible (gcj, kaffe, classpath, etc.), or they use Sun-licensed code (Blackdown, IBM, Apple, etc.). Something clearly makes it hard to re-implement Java, and that's probably both technical and legal. Whatever the specific reasons, it's a failure of Java as a general-purpose, standard programming language.

Whether Sun employees "bless" such a project or not doesn't matter: their opinion or public statements aren't legally binding. They know that it will take years until Harmony delivers anything, and Sun's legal team can still shut it down then.

Re:"blessing" doesn't matter

[Anonymous Coward]

"In reality, in 10 years, nobody has managed to create an interoperable, independent implementation of it."

Like TCP/IP right?

Re:"blessing" doesn't matter

[Anonymous Coward]

You appear to be in an alternate reality. There are hundreds of TCP/IP implementations for almost every platform in existence. Totally unlike Java.

Re:"blessing" doesn't matter

[Anonymous Coward]

A failure of Java or a failure of the respective software developers? The Java grammar is there for all to implement their own compiler and runtime. So what if it's too big for one person to do it alone, that's not Sun's problem or fault.

Re:"blessing" doesn't matter

[Jerf]

Something clearly makes it hard to re-implement Java

I'd like to tender a vote for "It's sheer Brobdingnagian size". Individually, each individual function of an API is something you could probably assign a college student to do, but taken together, to re-implement something like Java (which, like "Perl" or "Python" and unlike old-style "C" or ECMAScript, also implies a fairly sizable standard library) is just damned hard.

And as one lil' open source developer, I can't work up much excitement about re-implementing a language spec. (Full disclosure, I hate Java, but that statement is generally true; I can't think of any language I'd care to donate my time towards re-implementing.) I can't imagine this helps the developer pool. (Obviously this is not true of everyone, if you think I just claimed otherwise please learn to read what people say, not what you think they said. I'm just saying that I doubt this gets many people's blood pumping in a way that Yet Another Web Framework or YA MP3 Player seems to.)

Re:"blessing" doesn't matter

[Anonymous Coward]

Mono proves that if you have a funding source (Novell), it not that difficult to re-engineer a VM and class library.

No companies put money into independent because Sun is quite good at including them in the Java design process and community, so they have nothing to gain by going alone.

The problem isn't technical or legal, it's purely a lack of funding and desire.

Re:"blessing" doesn't matter

[Daniel Ellard]

Whatever the specific reasons, it's a failure of Java as a general-purpose, standard programming language.

Utter nonsense. Let's count the number of distinct implementations of Perl, Tcl, Ruby, Visual Basic...

Languages that are reimplemented frequently tend to be small, simple and appeal to language weenies (scheme, *ML) and/or there's money to be made.

The specs for Java have always been completely open. Anyone can reimplement it. The only restriction is that you can't call it Java unless it meets the spec (and proving that it meets the spec is, quite understandably, nontrivial because Java is a large, complex language).

If you like Java but want to change a few things, you're even free to do that, as long as you call it something else, like C#.

Re:"blessing" doesn't matter

[p3d0]

FYI, IBM has a cleanroom J2ME implementation, complete with JCL. IBM also has a cleanroom J2SE VM and JIT compiler.

Re:"blessing" doesn't matter

[Jay Carlson]

The specs for Java have always been completely open. Anyone can reimplement it. The only restriction is that you can't call it Java unless it meets the spec

This is a reasonable thing to do under trademark law; the Java mark would guarantee that customers would be receiving a system that met a particular quality standard.

So therefore Sun has no problem releasing all of its current implementation under a DFSG-free license, as "naji" (Not A J*va Implementation)? I could give you a nice shell script to fix any "java"-related trademarks in that case.

While you're at it, it would be a good idea to renounce any legal action against anyone practicing patents involved in any naji-related implementation, like IBM did in its blanket patent releases.

Oh right. That would mean Sun might give up control. Of anything. And since this is 2005, Sun has forgotten their spirit of 1988: "anybody can do this, but we're going to be the best."

I just wish Sun would try to lead rather than rule.

Re:"blessing" doesn't matter

[babble123]

It's interesting to note that there are several re-implementation of Python, at various levels of maturity (Jython,stackless Python,PyPy, and IronPython)

Re:"blessing" doesn't matter

[cahiha]

Utter nonsense. Let's count the number of distinct implementations of Perl, Tcl, Ruby, Visual Basic...

Perl, Tcl, Ruby, and Visual Basic are not standard programming languages, they are merely widely used.

The specs for Java have always been completely open. Anyone can reimplement it. The only restriction is that you can't call it Java unless it meets the spec

That's false, and repeating that lie often enough doesn't make it true; go and try downloading the Java specs. The official Java specs are proprietary to Sun Microsystems, as are most of the JCP specs. The contents of the specifications are covered by several patents. It is completely unclear at this point whether you can legally implement the Java specs, no matter what you call the result.

Re:"blessing" doesn't matter

[cahiha]

FYI, IBM has a cleanroom J2ME implementation, complete with JCL. IBM also has a cleanroom J2SE VM and JIT compiler.

IBM is a Java licensee, which permits them to do Java implementations. IBM is also a huge company and a key supporter of Java; it would be suicidal for Sun to sue them.

All the same, IBM has actually been trying to get a cleanroom Java implementation out, an implementation that would be unencumbered by Sun's intellectual property. Doing so has proved to be hard, and Sun has been putting up roadblocks.

My hope is that the combination of public pressure, of multiple attempts at open source implementations, and IBM will force Sun to back down. But there is nothing in principle to stop Sun from doing what SCO has been trying to do, and unlike SCO, Sun actually has a pretty good legal argument: their licenses are clear, their lawyers did a good job, and there is no question about who owns Java.

Re:"blessing" doesn't matter

[synthespian]

ML, which is a language for weenies, was used in the verification of C code for the Airbus A340 fly-by-wire system, which proves it has no use beyond Academia, and was never used in industrial applications.
http://www.astree.ens.fr/

Re:"blessing" doesn't matter

[Daniel Ellard]

go and try downloading the Java specs.

Ok, let's try java.sun.com. The JVM, the libraries, JINI, JNI, ... what exactly is missing? (JCP is the Java Community Process, which isn't a spec.)

Re:"blessing" doesn't matter

[Daniel Ellard]

I wrote that ML was popular with language weenies, not that it has no use beyond academia or in industrial applications. You act as though "language weenies" is a derogatory term, instead of a term that means "people who deeply care about programming languages." I guess you're not a weenie.

Re:"blessing" doesn't matter

[cahiha]

Ok, let's try java.sun.com. The JVM, the libraries, JINI, JNI, ... what exactly is missing?

Try downloading the specs and read the click-through licenses. They tell you that the information belongs to Sun and that you can't just go off and implement it in whatever way you like.

JCP is the Java Community Process, which isn't a spec.)

Yes, the JCP is the Java Community Process. The JCP produces specs. Those JCP specs are not freely implementable, they are proprietary to Sun.

Interesting name "Harmony"

[stevew]

What is interesting is that another project by this name got started during the QT is BAD days. Several developers started the "Harmony" project to replace the QT library with a GPL'd clone. Trolltech relicensing the library stopped this in it's tracks.

Odd how history DOES repeat itself ;-)

Re:Interesting name "Harmony"

[m50d]

I assumed that was deliberate. Maybe a subtle hint to Sun that they could make things a lot easier for everyone by dual-licensing java under GPL?

Re:Interesting name "Harmony"

[ccp]

Maybe a subtle hint to Sun

If that was subtle, I'd like to know what do you think is an unsubtle warning. Horse's head in the bed?

Of course the name was a warning. They're hitting Sun with the cluestick rather hard.

Cheers,

Carlos Cesar

Ummm...

[xeon4life]

Aren't there already like 3 or 4 FOSS Java implementations already?

Why is everyone getting hibbidy-jibbidy over this "proposed" implementation that was conceived on a forum?

Is it because a few high-standing developers are saying, "Oh yes, yes, this one, this one!"

I just don't get why this is different, other than maybe the licensing...

Re:Ummm...

[xeon4life]

Oh yeah, and I hate replying to myself, but if they really did build this from the ground up it would be the worst case of reinventing the wheel I'd have ever witnessed.

Re:Ummm...

[SharpFang]

More because this one is -finally- promised to be Java (the "real one") compatibile. All the free/open source java implementations are lacking more or less, usually more, falling short from being interesting to "real Java developers".
But why won't they pick one of existing platforms and change it to their needs, instead of starting from scratch, is beyond me.

Re:Ummm...

[wasabii]

They are lacking becuase they ARE NOT DONE.

It's disappointing to see the Apache guys start from 0% when we're already at 60%. It's going to take them years to catch up.

Re:Ummm...

[HvitRavn]

I'd just like to note that nothing has been decided yet. But most likely licensing issues will be overcome and Harmony will use existing projects to build the J2SE implementation. If so, that will be a win-win-win situation for alot of projects and people.

Not Embarassing

[mfh]

This is not embarassing. It's news that is good enough to repeat.

I can't wait for Java to be fully open source so I can gut it and re-release it how I would like to see it written. Can't wait.

Re:Not Embarassing

[Anonymous Coward]

Only that it is no news at all. Others have code (http://gcc.gnu.org/java/), Apache Harmony has just a proposal. Why not support real programmers instead of this paper tiger?

Re:Not Embarassing

[Anonymous Coward]

Not as embarassing as double posting a "dupe" message and getting them both rating informative!

Re:Not Embarassing

[Donny Smith]

> This is not embarassing. It's news that is good enough to repeat.

Hah!
It is irresponsible, iritiating, disgraceful and stupid, that's what it is.

>I can't wait for Java to be fully open source so I can gut it and re-release it how I would like to see it written.

Yeah, I'm sure you're skillful enough to issue a spelling-fixing patch.

Da Name

[Anonymous Coward]

Harmony, thats an interesting name. Its the same name that was used when QT was going to be replaced. Hum is Harmony going to be the name of the closed source killers?

Zzzzzz. Wake me up

[thammoud]

when this topic dies down. As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source? Sun along with the JCP has done an incredible job in advancing the platform. Java is the number one development environment for business applications. Bar none.

Why would someone encourage fragmentation and resource wasting ala KDE, Gnome and the gazzillions of Linux flavors is beyond me.

Sun, keep up the great stwerdess of the Java platform.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

Yeah, I agree totally. There's plenty of good JVMs around, these guys seem to re-invent the wheel all the time. Copying JDK1.5 is a massive undertaking, and it will probably never be as good as sun's version anyway. If Java was open sourced 5 years ago it would be in a mess today, in my opinion. Sun realized this, and that's why they kept control of it. On the other hand, I do like Eclipse though, but that was a commercial enterprise too actually.

And you are correct about Linux. It will never beat windows because people considering moving to Linux are then faced with the question "Which distribution?". This decision alone would scare off most people, but then you also have to choose which desktop to use! Crazy, and it will never attract the average person, only repulse them.
People would gladly pay the extra $200 to Bill for the certainty he provides. People expect windows to work. People will go to McDonalds and get the crappy burger they expect than try an unknown burger joint that is probably 10x better but might be a disappointment. Better the devil you know...

Open source people need to understand this.

Re:Zzzzzz. Wake me up

[Narchie Troll]

You have it backwards.

Free software will benefit from a free Java.

Re:Zzzzzz. Wake me up

[thegamerformelyknown]

...How isn't Java free right now?

Re:Zzzzzz. Wake me up

[dmaxwell]

Sun won't even take reasonable steps to allow distros to pre-install Java. Yes, it is simple for me to install Java after the fact. It is a legal licensing pain to distros to offer well integrated Java installs. That is one problem.

The other problem is that only platforms that are directly important to Sun or IBM get full featured Java environments. Java on PowerPC Linux is still substandard. IBM makes a JVM availiable but you have to jump through hoops even as an end user to get it and you still don't have a browser plugin. An Open Source Java would be available on just about all platforms with equal functionality.

Re:Zzzzzz. Wake me up

[MrDomino]

As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source?

Ever tried running Java on *BSD? It works to some extent, but it isn't pretty. Having an open implementation could mean that Java and Java Server Pages would become more widely accepted in servers running open systems.

Re:Zzzzzz. Wake me up

[Narchie Troll]

It isn't distributed under a free software license. Pretty simple. Any free software written for Java that makes use of features present in Sun's distribution but not in GCJ/Classpath is not truly free, since it depends on proprietary software to run.

See The Java Trap.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

The license is an arse and some Linux distributions won't include it on that basis alone. There is no binary distribution for the *BSDs because of some crap about not being able to distribute modified binaries AND the stupid thing needs Java to build the Java platform creating a chicken and egg situation on minority OSes.

Generally though, I'd rather they didn't OpenSource Java because that would concentrate more interest in Mono and there is no danger of accidently interacting with those Java heads who are so in for the money and not the love of it.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

Sun's JVM version is not available for use in all platforms and circumstances on which it might be desirable.

Re:Zzzzzz. Wake me up

[jyoull]

Cuz then those of us who use Java on Mac OS/X might finally have a faint chance of timely access to new Java releases, and JREs that aren't threaded thru with tripwires and platform-specific bugs that don't seem to be fixed very quickly.

Re:Zzzzzz. Wake me up

[electroniceric]

OK, this is the first substantive reason I've heard for spending years re-implementing what Sun and others have made reasonably available. Evidently Sun has not quite reconciled itself to playing in the OSS sandbox when it comes to Java, and they will probably have the jitters for a while yet while they figure out what exactly they're trying to sell.

But aside from licensing holy wars, if the problem is that Java is hard to distribute w/ OSS, shouldn't the solution figure out how to pressure Sun or IBM into distributing Java with OSS - or shucks - just make the install process thoroughly transparent? This is like saying: there's no OS with using the Mozilla license! I'm gonna implement a new OSS, that walks and quacks like Linux, but is completely licensed under MPL rather than GPL! Where's the beef here?

Re:Zzzzzz. Wake me up

[mav[LAG]]

Java is the number one development environment for business applications. Bar none.

Bahahaha. Number one in what? Being slow? Being broken? Being inconsistent? Being verbose? Being a nightmare for sysadmins to manage?

No doubt Java fanboys will mod me down for trolling but I don't really care - the above has been my real world experience of it. Any Lisp or Python programmer worth his salt can code circles around anything written in Java, and it will be written in a quarter of the time and run twice as fast. But of course PHBs have a far better grasp of what language to use than programmers do for some reason.

Re:Zzzzzz. Wake me up

[thegamerformelyknown]

Hmm, yes, good point. I wasn't thinking of "free" in that context.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

LOL. Yes. Lisp. How about APL? or some other equally retarded language only mathematicians, washed-up AI people & other various people who never got past just being a programmer (vs designer vs software engineer) could love...

Because LISP scales *so* well. And has all kinds of useful features like ... eval!!!!

And it's so fast *every* systems project that is worth anything is written in LISP!

Yes! the way of the future! RPN & expressing yourself in syntax trees! What next? Graphical programming??? You just draw the syntax tree. So now you can have a *picture* that no one else understands instead of just a program... *rolls eyes*

Re:Zzzzzz. Wake me up

[Wesley Felter]

if the problem is that Java is hard to distribute w/ OSS, shouldn't the solution figure out how to pressure Sun or IBM into distributing Java with OSS?

People have been working on that problem for almost ten years, and no progress has been made. It's time for a different approach.

Re:Zzzzzz. Wake me up

[ignorant_coward]

Number one in the low number of exploits. How many J2ME worms are out there? How many Java applet hijacks are reported each year? Close to zero, if not zero.

Also, there are plenty of benchmarks showing Java is as fast or faster than C and C++ on large datasets and long-running applications, when the environment initialization isn't a hit on performance.

Java isn't perfect, but it is so complete that it would be easier to use than most alternatives. While Python is certainly gaining traction, Lisp quickly becomes non-portable once the project is large enough (contrary to popular belief).

Re:Zzzzzz. Wake me up

[mav[LAG]]

LOL. Yes. Lisp. How about APL? or some other equally retarded language only mathematicians, washed-up AI people & other various people who never got past just being a programmer (vs designer vs software engineer) could love...

I love people who don't know what on earth they're talking about. Lisp has little to do with AI except as an accident of history. Today people use Lisp to write 3D games for the Playstation, complex business applcations, robot controllers for NASA and just about anything in between.

Because LISP scales *so* well.

Well spotted - it does. Steel Bank Common Lisp on my AMD64 compiles to wickedly fast native x86_64 code - actually faster than gcc in some cases.

And has all kinds of useful features like ... eval!!!!

From http://www.flownet.com/gat/jpl-lisp.html:

(Debugging a program running on a $100M piece of hardware that is 100 million miles away is an interesting experience. Having a read-eval-print loop running on the spacecraft proved invaluable in finding and fixing the problem. The story of the Remote Agent bug is an interesting one in and of itself.)

Yes! the way of the future! RPN & expressing yourself in syntax trees!

Yeah Lisp is the way of the future actually - heh. All other actively-developed languages have only recently added things that have been in Lisp for decades: closures, GC, macros et al. Lisp isn't RPN by the way which shows me how much you really know about Lisp. And you don't typcially write code in syntax trees - you write a domain-specific language in Lisp and then write your problem in that. Reply as yourself if you feel the need - I'd like a good laugh at your previous cluelessness.

Re:Zzzzzz. Wake me up

[argent]

As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source?

It's not whether it's open source or not, it's whether it can be shipped with an open source OS or not. Right now the biggest problem Java has is that when you get your BSD or Linux system set up, you can just click buttons in the installer and you're good to go.

Unless the packages depend on Java, then you have to download Java from Sun's website (which can be really annoying when you're behind a hard firewall and you weren't expecting this) and THEN go back and install them.

So I don't care if Java is the #1 development environment for ANYTHING, right now it's just a royal pain in the ass.

If Sun wants to keep people from reverse-engineering and fragmenting Java, they need to figure out how to fix their licensing so people can just treat it as something else that ships on their Fedora or FreeBSD or whatever CDs. I don't care how they do it, or what face-saving games they have to pull, I just want to stop getting that "oh no not again" feeling when I see that some program I want to use is based on Java.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

Why would someone encourage fragmentation and resource wasting ala KDE, Gnome and the gazzillions of Linux flavors is beyond me.

Encourage fragmentation? That is what Sun is doing by not making their own implementation open-source. Would Harmony exist if Sun's implementation was available under acceptable license terms? Does Sun's excuse of "avoiding fragmentation" justify keeping it closed when it actually promotes fragmentation?

Not that I think multiple independent implementations are a bad thing. EGCS did wonders for GCC, Apache grew out of NCSA's httpd, and Linux came from UNIX.

Re:Zzzzzz. Wake me up

[iggymanz]

To date, Sun has only lost money on java. They are also losing server market share, and they still push their proprietary operating system (the current version of which is 30% vaporware, btw) which costs over $200M a year to develop. What happens if Sun drops java as too expensive, or Sun dies, or is purchased by a hardware vendor which has no interest in pouring cubic yards of cash into a language? Be nice if Java had a growing worldwide non-MS centric force behind it such as open source.

Re:Zzzzzz. Wake me up

[junkgui]

...foreach... thats your best argument... because the other two just arnt true... public class Speaker { public void callTalk(Object o) { o.getClass().getMethod("talk", new Object[0]).invoke(); } } And threads arnt broken... except that they are still backward compatible with java 1... where they were broken but he methods are depricated and have been for many many years.

Re:Zzzzzz. Wake me up

[Ars-Fartsica]

Also, there are plenty of benchmarks showing Java is as fast or faster than C and C++ on large datasets and long-running applications, when the environment initialization isn't a hit on performance.pAlso known as contrived tests.

Re:Zzzzzz. Wake me up

[radish]

You can do both of the "impossible" things in Java, of course you can. There just isn't a one-keyword syntatic sugar. Is that really the best advantage you can come up with for some other language, that it has more sugars? Please. I can do a foreach in my IDE as simply as typing itco[tab]. I can do your second example in pretty much the same way. 5 keystrokes. Now, your problem again?

Re:Zzzzzz. Wake me up

[mcc]

>> Also, there are plenty of benchmarks showing Java is as fast or faster than C and C++ on large datasets and long-running applications, when the environment initialization isn't a hit on performance.

> Also known as contrived tests.

Actually it's called "J2EE".

Re:Zzzzzz. Wake me up

[Furry Ice]

The problem is that they won't get equal functionality in an open source VM, either. Writing a good JIT is a lot of work, and it needs to be done for every platform. Obscure platforms just won't get the same level of attention. It will probably be better than Sun's support, but still only the most popular platforms will get top-notch support. Unfortunately, it's just the way it works. Look at how frequently Linus breaks the compile for non-x86 platforms!

Re:Zzzzzz. Wake me up

[mr3038]

This is like saying: there's no OS with using the Mozilla license! I'm gonna implement a new OSS, that walks and quacks like Linux, but is completely licensed under MPL rather than GPL!

There's one critical difference. Linux is licensed under GPL plus a clause that allows running non-GPL software on the system. That allows users to run software with GPL incompatible licenses already so there's no need to rewrite MPL (or any other license) compatible variant. Linux can be also freely distributed and its modified versions can also be re-distributed as long as the modifications are released under GPL license. On the contrary, Java license doesn't even allow distribution of verbatim copy of VM, much less re-distribution of derivative works. There's no contest here.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

"Python with a bit of C handles it happily though."

Java with a bit of C handles it happily though.

"It's only just getting stuff that Python and LISP programmers take for granted."

Foreach is syntactic sugar. I can't say either way on 'latent typing' (don't know). Java was a fully threaded language from day one. If you can show how it is broken even after a decade, you should consider telling Sun about it, cause you are clearly the smartest person in the universe.

"Common Lisp on the other hand has a standard which has been around for nearly fifteen years. It is highly portable."

Yeah, but what about GUI toolkits? What about doing anything useful with POSIX interfaces?

Lisp is great, but building big and useful apps with it requires buying into proprietary frameworks from the major Lisp vendors.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

Hmm. I'd like to see a corporate manage 500,000 lines of lisp code. People complain that my Java code is messy and unreadable... wait till they see my lisp code. I will redefine "unreadable"

Re:Zzzzzz. Wake me up

[divec]

As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source?

Here's one real example:

• I want to distribute OpenOffice.org 2 CDs to people in rural communities without broadband.
• Large chunks of OOo2 depend on Sun's JVM.
• I cannot legally redistribute Sun's JVM.
• Therefore, to get a fully functional OOo 2, they have to spend hours downloading, and the install process is far more complicated.
• Therefore, more people stick with Microsoft Office.

This exact issue kills off large parts of the desktop market for Java at the moment - end-users will go for a solution which is not such a pain to install, rather than one written in Java, if they have a choice.

Re:Zzzzzz. Wake me up

[drew]

As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source?

As I've never been a Java developer and most likely never will be, I don't much care what license sun distributes the java development tools under. However, as a person who occasionally uses applications developed in java, it would be nice if there was at least a freely redistrbutable if not open source runtime, so that it wouldn't be such a pain as an end user.

Re:Zzzzzz. Wake me up

[jtpalinmajere]

Ever tried running Java on *BSD? It works to some extent, but it isn't pretty. Having an open implementation could mean that Java and Java Server Pages would become more widely accepted in servers running open systems.

To be honest, this is the case on *every* platform I've developed with / used Java on... but that is just my humble opinion. There are plenty of projects where I and my colleages were restricted to Java when we could have done a better and faster job using C or Python. The only thing that makes Java even remotely nice IMO is its ability to keep stupid programmers from doing things *too* stupidly... but these arbitrary barriers often only serve to get in the way once you've graduated from programmer pre-school.

In any case, I'd be more interested in going the C# route of stupid coder friendly languages... being that it essentially tries to 'fix' just about everything that is problematic in Java. I'm not saying its a perfect solution... it certainly has plenty of revisions or forks to go before we see anything resembling a holy grail of OO noob friendly programming languages. ...Even with an immature runtime as Mono to run C# apps on... it at least offers open-source portability to X platform.

Re:Zzzzzz. Wake me up

[malachid69]

Actually, I am doing that now. I am currently running the Linux JVM on top of the Linux compatability layer. No difficulties getting it installed or running.

Re:Zzzzzz. Wake me up

[FriedTurkey]

In any case, I'd be more interested in going the C# route of stupid coder friendly languages

C# actually puts in a lot of stupid coder pitfalls back in trying to 'fix' Java.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

If the Lisp version is 500 kloc, the Java version would probably be 5 Mloc, with any recurring control structures and mixin class features dutifully copied and pasted everywhere. The point of Lisp is to let you implement the language you need to easily maintain your system!

Re:Zzzzzz. Wake me up

[Anonymous Coward]

Lisp is actually pretty difficult to optimize, Python nearly imposible. Static typing and simple link-time method dispatch decisions are very important hints for the code generator.

Re:Zzzzzz. Wake me up

[jtpalinmajere]

C# actually puts in a lot of stupid coder pitfalls back in trying to 'fix' Java.

All too true, but that's what I'm lauding in part. I don't mean to sound arrogant, but I don't exactly consider myself and many of my colleagues to be in the "stupid coder" crowd. I don't even particularly prefer using C# at all... only that I prefer it to the likes of Java when forced to adopt a business's risk mitigation attempt to make the development process as stupid-proof as possible... not to mention it technically has an open source development platform (Mono) that, while not mature, is surprisingly functional and extensible when needed.

I personally think that Java is a great learning tool... something that our intro to programming and a handful of other introductory classes in college should adopt and use. Past that, however, I can only compare it to the likes of Visual Basic... almost as if it is only a language that only those who haven't been weaned (or don't want to be) off it and move onto more powerful languages (albeit, languages requiring more developer responsibility and vigilance).

In the end, I think the approach of creating stupid-proofing mechanisms into our languages is flawed... we'll only create better (worse) stupid people. Instead, take the correct (albeit far more difficult) course of pruning the development community of stupidity. Greater education rigor including principles and disciplines would do much more than the majority of our universities' current Java shop curriculums.

Re:Zzzzzz. Wake me up

[FriedTurkey]

I went from a C/C++ background to Java. I really have no inclination to go back. Java allows me to get stuff done quicker. Do I miss messing with pointers? Uh no.

No Java was taught when I was in college when I went. Going to Java in the late 90's was considered "cool". I guess now since it is taught in college it kinda has that "establishment" feel. Just because there is a population who learned Java in college and is now writing crappy code doesn't make Java the problem. If they teach Ruby and Python in school it just moves the problem to another language. Maybe the problem isn't Java but the education system?

Re:Zzzzzz. Wake me up

[jtpalinmajere]

Fair point... though I'm inclined that the answer lies somewhere along the lines of "all of the above". Having universities become a C shop or Python shop (it *would* be nice if it was a language independent *theory* and *discipline* shop... but that's dreaming too much) still doesn't change my baseline qualms with Java, though.

Re:Zzzzzz. Wake me up

[Anonymous Coward]

I.e. about as good as running Windows software in Wine... That doesn't make Windows software cross-platform either.

Re:Zzzzzz. Wake me up

[Beuno]

I see your point of people sticking to office, but I don't think it makes sense they would purchase a license so expensive and not have any type of broadband connection. Unless they where buying a pirated copy, then we wouldn't care a rat's ass about Sun's licence. Don't get me wrong, I agree 100% that it's stupid not to be able to distribute JVM that way, but maybe then OO should't depend on Java at all.

standard java?

[dukerobinson]

Is the reason that SUN won't just GPL the regular Java because they want to have final say on what is the gold standard of java?

Re:standard java?

[m50d]

That's what they claim. But they could QPL it, that way there wouldn't be fragmentation except for really really important things (which sun would probably add in quickly themselves). And they keep control of the Java trademark, so any non-standard Java can't call itself Java. IMO they're just making excuses.

Re:standard java?

[ClosedSource]

As Spock once said in a Star Trek novel "A difference that makes no difference is no difference".

There's really little value in opening Java unless you need to change it in some significant way that the current process doesn't allow.

To the extent that an open Java is useful to others, it hurts Sun's power over it. To the extent it doesn't hurt Sun's power over it, it is not useful to others.

Re:standard java?

[Bloater]

If I need to use a Java program, and the JVM is not open, then I lack control over a significant portion of my computer and work. This is a bad thing and an immoral thing. I will welcome a free-software JVM.

Re:standard java?

[team99parody]

Of course you're trolling, because there is nothing about the GPL that would prevent Sun from defining the standard.

For an obvious example, consider C and C++, which are both supported by gcc (a GPL'd work), yet the ISO committees still are able to define the standards.

Re:standard java?

[petermgreen]

the benifits of a free java

1: people can fix bugs DIRECTLY and distribute those fixes
2: distros can distribute java freely so it works out of the box
3: people can port it to new platforms freely

and those are just the practical reasons. There are people who prefer free software because they like to have the freedom even though they don't use it themselves.

Re:standard java?

[ClosedSource]

You and Bloater missed my point. The post I was responding to was discussing the issue from Sun's point of view, not users and developers.

To paraphrase an old Woody Allen joke about sex and relate it to OSS:

Q: Does OSS reduce the power of proprietary companies?

A: Yes, if you're doing right.

Re:standard java?

[m50d]

Making it QPL would allow it to be included in Debian and other linux distros that haven't bothered to negotiate a license to redistribute from sun. This would be a substantial benefit to them, their users, and also to Sun because it would make it easier for end users to use Java. At the same time I don't see how it would harm Sun's power over it.

Re:standard java?

[ClosedSource]

"At the same time I don't see how it would harm Sun's power over it."

How could it not? If the license puts any restrictions or requires any additional actions on or by Sun with respect to Java, it lessens Sun's power over it.

Re:standard java?

[malachid69]

Actually, they don't have final say. Apache has the same vote that Sun does on the JCP Executive Committee. All Sun has is a permanent seat on the board.

Re:standard java?

[m50d]

How could it not? That's a nonsense question. It won't harm sun's power over java unless it deliberately does so.

It doesn't put any restrictions on Sun, it's their license, they hold the copyright anyway so they're not bound by it.

big deal

[Anonymous Coward]

And it's causing quite a stir in the Java community, especially since respected Sun frontmen Tim Bray, Simon Phipps, and Graham Hamilton have given the project their blessing.

That's kind of like getting Toto to give a "blessing" for the rights to the Wizard of Oz.

Is something going on behind the scenes ?

[johnhennessy]

Has Sun donated to the Apache foundation before (XML parsers ??)

If they were trying to release source code as a seperate effort wouldn't the Apache group be a likely outlet ?

Re:Is something going on behind the scenes ?

[benjamindees]

Is something going on behind the scenes ?

Pay no attention to that monkey... er, man... er, monkey-boy behind the curtain.

Re:This is great news...

[Anonymous Coward]

You know you can just rename "plugin" directory of Acrobat reader to make it load waaaay faster.

Yeah

[Anonymous Coward]

We should instead go with C# and the CLR, which are much more free and open because all implementations besides the Microsoft one, such as Mono, are equally unauthorized and under no obligations whatsoever to work

Re: Blackdown's builds

[codergeek42]

Blackdown's JRE/JDK builds aren't open-source or freely distributable.

Re:Death to Apache

[Anonymous Coward]

Bill? Is that you?

Re:Death to Apache

[Anonymous Coward]

I like your ranting style!

There is for sure much to be desired in the Java world, but the Harmony vapourware is not what is needed.

OpenOffice.org

[codergeek42]

Perhaps a usable F/OSS Java implementation would quelch the OpenOffice.org 2.0 and Java issues...

Re:OpenOffice.org

[m50d]

Absolutely, except the problem is they use internal sun-only classes (sun.* packages), so no other JVM, even a completely standard-compliant one, can run OOo.

What about patents?

[unoengborg]

Didn't Sun and Kodak have some patent dispute over Java a while ago where Kodak won. What risks would this Apache projects involve with respect to Kodak patents?

Re:What about patents?

[Anonymous Coward]

You mean Sun really paid Kodak to sue anyone making a non-Sun licensed Java.

Re:What about patents?

[the eric conspiracy]

What risks would this Apache projects involve with respect to Kodak patents?

The Kodak patent is so broad that it could be used to sue anyone using an object oriented programming language. It is not Java specific. Sun settled with a $92 million payment, Microsoft has taken a license.

Unfortunately it seems to have survived a court test, so it will take a lot to get it declared ivalid. However most people believe that it should be because of prior art going back to the days of Simula.

Harmony Project

[Anonymous Coward]

The Harmony Project was the name of a project to implement a Free QT clone back in the early days of KDE.

Working Code vs. Proposals

[Anonymous Coward]

GCJ http://gcc.gnu.org/java/ has working code. The Apache guys just have a proposal with some box diagram and some "blessing". Ho, ho, ho, which one is more useful when it comes to writing a Java program?

Why on earth can't the Apache guys bend over and support GCJ (with code, not with words)? It looks as if Harmony is more an ego me-too thing, than anything else.

Re:Death to Apache

[ShinmaWa]

Thank you for copy-and-pasting, verbatim, this blog entry

The link would have been sufficient (oh wait, the link was given -- in the first post!)

Getting ridiculous

[Exaton]

Zonk, you are an embarassement, an icon of shame for this website.

You've been around long enough now. Time to start quickly checking /. archives before accepting news submissions.

I hope you're really, really ashamed of the multiple editorial failures you're responsible for. Time to step up and do something about it, man ! No need to follow bad examples !

And I choose not to go AC to say that.

Re:Getting ridiculous

[Anonymous Coward]

Please report to siberia.slashdot.org for decommisioning. You have spoken out for the last time.

wow, that was totally not funny. *checks the "post anonymously button*

sorry

Why Another Open Source Java?

[Anonymous Coward]

Because OpenOffice won't build (unmodified) under GCJ, and open source folks don't want to rely on proprietary Java in OpenOffice.

I see a case of "redundancy"

[bogaboga]

Tell me please: Why should I, as an end user, download and use Apache's Harmony instead of using SUN's "real" Java. They are both "free" for me to obtain. In short, I wouldn't. I also believe that for software that requires Java, application developers will continue to direct users to http://java.sun.com/j2se/1.4.2/download.html. So there you go!

Re:I see a case of "redundancy"

[k98sven]

As an end user? No real reason why you should download one over the other. Ideally one will do as well as the other.

However, if you're using Linux (most distros don't accept Sun's licensing) you won't have to do that, since a FOSS distro wouldn't have the same terms.

If you're using an operating system which Sun's java doesn't support, you're plain out of luck. A FOSS JVM is more likely to be ported. And you can do it yourself if bad comes to worse.

Lastly, having a second implementation avoids vendor lock-in. Today there is no other implementation of Java than Sun's. (Apple, IBM, Blackdown and all that are based off Sun's code)

Re:I see a case of "redundancy"

[damiangerous]

Tell me please: Why should I, as an end user, download and use Apache's Harmony instead of using SUN's "real" Java.

Scenario 1: You as an end user wouldn't have to. It would be included with your operating system, which is not currently the case. If you have to seek one out you would probably seek out Sun's, but if your Linux distro came with Apache's you would just use it unless it was unsuitable in some way.

Scenario 2: You develop Java apps. Right now you have to direct end users to another website (Sun's) and follow instructions found there to download and install Java. You could instead offer an install package that already includes Harmony.

Scenario 3: You want to use Java on an unsupported machine. Right now you don't really have many options for running Java apps on PocketPCs, for example (Like my Jornada 568). An Open Source JVM would almost certainly be ported a wide variety of platforms (considering how many platforms have reimplementations like Waba attempted for them the demand is obviously there).

Re:I see a case of "redundancy"

[SharpFang]

As end user, you aren't interested in Java per se, you are interested in applications. So, you want to run an application written in Java.

Currently, the download is:

"[download NOW] --- This program requires a Java Virtual Machine. You can obtain one from [SUN website]" and then several steps of separate install to get Sun J2SE working.

This would look now like this:

"This program requires a Java Virtual Machine. If you already have one, download [standalone version]. If you don't, or aren't sure, download [version bundled with Harmony JVM]". One simple choice, maybe one extra click-through page of the installer in the "expert" version, you have the app plus java installed, app ready to run.

Re:I see a case of "redundancy"

[argent]

Why should I, as an end user, download and use Apache's Harmony instead of using SUN's "real" Java.

As an end user... you won't have to go to anyone's website and accept a clickthrough license to download Harmony. It'll be there on your system, like Perl and bash are.

Re:I see a case of "redundancy"

[JohnnyCannuk]

Scenario 1: You as an end user wouldn't have to. It would be included with your operating system, which is not currently the case. If you have to seek one out you would probably seek out Sun's, but if your Linux distro came with Apache's you would just use it unless it was unsuitable in some way.

This would make sense except that Dell and HP already sell Windows machines with Sun's latest JVM installed by default. So more than likely an "end user" already has Java installed. If its linux well, the same goes - my Suse 9.1 came with both Sun's and IBM's 1.4.x JDK installed. Harmony would solve the problem for the more strict distributions like Gentoo and Debian and probably for *BSD. I don't think this a particularly good reason.

If a user is smart enough to download and install the app they want, they are smart enough to download the JVM. And they should only have to do it once, since the automatic update feature has been part of the JRE since 1.4.1.

Scenario 2: You develop Java apps. Right now you have to direct end users to another website (Sun's) and follow instructions found there to download and install Java. You could instead offer an install package that already includes Harmony.

Try re-reading the licence. The JRE is freely redistributal. I have seen a great many software package installs that come in versions with or without the JRE included. Most developers assume that if you are downloading a Java app, you already have Java. If they don't, they would rather let you get the latest verion from the source rather than maintain their own version. You know, "the man with two watches never knows what time it is."

Scenario 3: You want to use Java on an unsupported machine. Right now you don't really have many options for running Java apps on PocketPCs, for example (Like my Jornada 568). An Open Source JVM would almost certainly be ported a wide variety of platforms (considering how many platforms have reimplementations like Waba attempted for them the demand is obviously there).

Now this is a good reason for an Open Source Java. I would love to have a Java I could tweak or compile for an iPaq, etc.

Open Source Java would speed adoption on a wider variety of "non-standard" platforms from a technical standpoint. This will make Java more ubiquitous (and probably better and more efficient as well). It will also satisfy the licencing zealots, having the same effect.

I think Harmony and an open source Java are a great step forward, but not for the two reasons mentioned.

Re:I see a case of "redundancy"

[zander]

• Scenario 2: [snip]
  Try re-reading the licence. The JRE is freely redistributal. I have seen a great many software package installs that come in versions with or without the JRE included.

When an application is 200Kb it suddenly grows about 7Mb in size to include the JRE. I know of quite some people that illegally strip the JRE of parts they don't need. Its a bit silly to have to do something that makes your app shippable illegally.
Next to that; including the JRE in your package does not make sense on systems that have good installation-managers that handle dependencies for you. So it only makes sense on Windows or when the user downloads the package, which surely is a very tiny part of the userbase.

Cheers!

Re:Possibly poor foresight.

[junkgui]

Apache alread is part of the JCP...

This is GREAT!!

[Anonymous Coward]

I can't wait to see an open source version of java. It will be wonderful because:

1. It will break with the Java standard and do things "better", IE be incompatible.
2. It will run on HURD
3. RMS will give it his blessing
4. It will be choked full of bugs, since it's written from scratch.
5. Will cost nothing, which is much better than the cost of the JVM's and tools such as Eclipse!

Oh, open source Java is the cure for all our problems!

Re:This is GREAT!!

[Anonymous Coward]

First let me tell you, you won the prize for the dumbest comment I have read today. Congratulations!

1. It will break with the Java standard and do things "better", IE be incompatible.

Ummm, their goal is TO BE compatable with J2SE 1.5.

2. It will run on HURD

And Linux/PPC, FreeBSD, OpenBSD, NetBSD, Linux/Sparc, Linux/Itanium, SCO Unixware/Openserver (shudders), And numerous other Operating Systems.

3. RMS will give it his blessing

Ummmm, I really highly doubt that this is one of the Apache Group's motivations.

4. It will be choked full of bugs, since it's written from scratch.

Most likely, considering it will be in Alpha and Beta first.

5. Will cost nothing, which is much better than the cost of the JVM's and tools such as Eclipse!

This also is not one of their motivations.

What happens if Sun goes under? Or decides Java is too expensive to maintain and starts charging for it? Hmmmm? I suppose we are supposed to wait for that to happen AND THEN start working on an open jvm?

Re:This is GREAT!!

[Anonymous Coward]

You obviously know nothing about the Java Community Process, or the fact that Sun does not wield "control" over Java, they are merely a member in the Java community. Should Sun die, and they probably will, Java will live on.

Java cannot be open source because it will then lose its cross platform compatibility. This compatability is one of its main advantages, but also a total pain in the ass. Many people dont care about the cross platform features and would trade some of it for things like better floating point & scientific features. This type of thing would lead to different flavors of java, and that can not happen.

To keep the compatability Java can never be open source.

Re:Not Embarassing (OT)

[Anonymous Coward]

That's the lowest UID I've ever seen post :P

IBM connection

[wrmrxxx]

What makes this slightly interesting is the IBM connection:

Geir Magnusson Jr. is a lead in the proposed Har mony Project

Geir Magnusson Jr. is from Gluecode, which IBM has acquired.

If it weren't for that, I'd just say "yeah, whatever - it's just another JVM implementation."

Please Add Decimal Math (fixed-point) as a builtin

[Anonymous Coward]

The title says it all.

More Convenient Installation

[jack_csk]

Does it mean a bit more convenient way to install J2SE on my Gentoo and FreeBSD boxes, without worrying about downloading files manually (especially during software upgrades) or having licenses being revoked?

I, for one, am excited about this new Harmony project.

Re:Possibly poor foresight.

[Anonymous Coward]

The GPL is incompatible with the Apache licence, not visa-versa. And there's much more Java code under the Apache licence than under the GPL.

Sun already made JVM source available

[Anonymous Coward]

If they were trying to release source code as a seperate effort wouldn't the Apache group be a likely outlet ?

Sun already released source code of Java 1.5:
http://www.sun.com/software/communitysource/j2se/j ava2/download.xml

It's under Sun Community Source License or under Java Research License. Obviously, none of these is GPL compatible.

Re:Possibly poor foresight.

[mcc]

Apache alread is part of the JCP...

Okay.

How actively do they participate in it, at current? What degree of influence or voice do they have? Is the creation of the Harmony project likely to change either of these things?

Fending off the fundamentalists...

[aquarian]

As a Java developer for the past 7 years, can someone remind me as to how Java will benefit from being open source?

By not giving excuses to rabid, open-source fundamentalist freaks to attack every worthwhile project that uses Java?

Re:Please Add Decimal Math (fixed-point) as a buil

[Anonymous Coward]

why?

If it's fixed point then just do all your maths on an int or long and divide by the appropriate power of ten when displaying.

That's all that would be happening internally anyway.

You need a reason?!

[Anonymous Coward]

Have a look at this : http://www.sun.com/

I don't get it...

[mark-t]

People wine at Sun "Open Java... Open Java... Open Java...".

And when they finally look at doing so, all I see people saying are things like "We already have GJC, you fuckers... we don't need you anymore".

I just don't get it.

Re:I don't get it...

[argent]

I don't care if they open Java at this point... I just want them to allow verbatim redistribution of their own j2se because I'm getting really tired of having to download the right version of Java separately from Sun's site when I need to set up Java-based packages on Open Source UNIX.

That's what's more or less burned out whatever interest I ever had in Java. It's just another checkbox to slow down setting up a production system.

Re:I don't get it...

[Wesley Felter]

People wine at Sun "Open Java... Open Java... Open Java...".

Those people are whiners anyway; they have no credibility.

And when they finally look at doing so...

You didn't understand the article correctly. Sun is still not planning to open source their VM.

all I see people saying are things like "We already have GJC, you fuckers... we don't need you anymore".

One benefit of open source is that it makes users independent from vendors. Not needing Sun is very important for some people.

Apache made the political mistake of implying that they'll throw away GCJ etc. and start from scratch, which understandably has people in shock at the sheer hubris of it. (Of course, it doesn't matter whether that implication is true or not...)

Re:I don't get it...

[m50d]

Because this is not a complete open java implementation being released, it's a new project starting from scratch with zero code. Which seems pointless.

Why ask why?

[rdean400]

The fact of the matter is that there are certain of Sun's terms that don't suit the business model or philosophy of other parties. Harmony will provide a way to provide a version of Java that is compatible with those interests.

I'm a power user that doesn't need support for my operating system, so there's no reason to buy the boxed versions other than to provide financial support to the vendor or to acquire software the distribution is prohibited from providing for free download due to licensing restrictions -- like Java. It doesn't make sense to pay to get something free.

Kaffe and GCJ don't cut it because they are not completely compatible to the spec and their performance is woeful compared to Sun's JVM, let alone JRockit or J9. This will provide a version of Java that distro vendors and others can bundle with their products on terms compatible with their licenses, business models, or other philosophical beliefs.

Re:Online comment

[Anonymous Coward]

Who modded this insightful? This is a troll.

Legality?

[adolfojp]

Project harmony appears to be to Java what mono is to c#.

Anti Micsrosoft zealots bash Mono everytime they get. I'd love to read their oppinion towards Harmony.

Adolfo

Haha, nice work mods.

[Some Random Username]

If you aren't going to read the link, don't moderate the post. Its a troll post from a troll weblog.

Re:Please Add Decimal Math (fixed-point) as a buil

[Anonymous Coward]

Not the same AC, but it would be a nice builtin feature because then JDBC and other interfaces could understand it.

The apache 2 license is not free.

[Some Random Username]

They put in a bungled mess of patent nonsense, despite software licenses being for copyright law, not patent law. This mess is worded such that if you use software licensed under this license, you agree never to sue them, even if they say, steal your GPL code and stick it in a project of theirs. If you do sue them, then you aren't allowed to use the software anymore, and they can sue you for using it. It may not have been intended to facilitate code theft, but it can still be used that way, and despite being warned of this issue before the license was finalized, the apache people went ahead with the broken version anyhow.

From the department of redundancy dept

[BovineSpirit]

Is the jr in Geir Magnusson Jr not redundant?

Re:From the department of redundancy dept

[Anonymous Coward]

Yes, it is not. While Iceland still uses the naming convention of the baby's surname being the name of the father with a -son or -dottir suffix, the other Scandanavian countries have moved to the more familiar (to westerners) style of surnames. I think you will find that Geir's parents are Mr and Mrs Magnusson.

Re:From the department of redundancy dept

[Anonymous Coward]

No, the "Jr." is not redundant:

Can you tell from the name of "G M-son" that he is the son of "G M-son" without appending "Jr." ?

No, you cannot, because:

Assuming the M-son is used in the Scandinavian way, he is the son of "M X-son", not the son of "G M-son", which is inconsistent with the use of "Jr.", therefore the M-son is not used in the Scandinavian way.

As the M-son is not used in the Scandinavian way, and one has no idea of his father's first name, "Jr." is needed to show that it is the same as his.

Q.E.D.

Of course, anyone writing to a "News for Nerds" websites should be able to work that out.

Apart from Zonk of course.

And maybe CrmdTocac.

And Timothy.

And Timothy.

Miguel's take on Harmony

[Anonymous Coward]

here

Re:Miguel's take on Harmony

[/ASCII]

Yeah, ask the main PR man for a competing product what he thinks of more competition. You're bound to get some unbiased answers.

Personally I don't see how anyone can care about .Net vs. Java, they are so close to identical it's not even funny. As to the implementation, neither GCJ, Kaffe or Mono is 'there' yet in terms of quality, at least not in my experience. If we just leave people to do their hacking, evolution will weed out the bad implementations and we'll have the best tools available.

Re:Miguel's take on Harmony

[Anonymous Coward]

Mono works great and has dozens of high quality applications. Why trust some unknown on Slashdot when you can see the results for yourself. Calling Miguel a "PR man for a competing product" just shows your ignorance. He's the guy who started GNOME, wrote GNUmeric and generally got people as excited about the desktop side of UNIX as Stallman and Torvalds did for the operating system.

Re:Miguel's take on Harmony

[Anonymous Coward]

He's the guy who started GNOME, wrote GNUmeric

Yeah, he's also the guy that seems to think Microsoft is actually doing some good things (not) and that C#/.Net need to be the future of Linux.

I (among others) generally disagree with that posture.

and generally got people as excited about the desktop side of UNIX as Stallman and Torvalds did for the operating system.

First of all, I'm pretty sure Steve Jobs and NeXT/Apple, or even Scott McNealy and Sun, did much more to get people excited about the "desktop side of UNIX" than Miguel ever did.

Beyond that, I feel his "vision" is "flawed".

Hope that helped...

Re:Miguel's take on Harmony

[trixy_1086]

I saw "Miguel" and "evolution" in the same sentence, and was promptly confused about what an email client had to do with .Net or Java. I should probably go review the english language.

Re:Miguel's take on Harmony

[synthespian]

OK, reality check now people! Linux is a competitor to Sun. It competes with Solaris. Sun would not want RedHat (main Linux distro cited by Sun people as "the enemy") to have Java.

RedHat probably sees this and is probably behind Havoc Pennington's comments last week about why "Red Hat not shipping Mono is currently a can't rather than a won't. Making it worse, we are not able to spell out all the facts on why we can't."

http://galaxy.osnews.com/email.php?blog_id=973

Which means RedHat will make the technological mistake of investing too much in Java, in order to get at Solaris, along with Harmony project folks who in all naiveté believe in Sun "blessing" Java. All the while, Mono just sits there, waiting to be used more...

Re:Miguel's take on Harmony

[CodeArtisan]

OK, reality check now people! Linux is a competitor to Sun. It competes with Solaris. Sun would not want RedHat (main Linux distro cited by Sun people as "the enemy") to have Java.

Competitor as in, Sun offers linux products ?

So what changed in the license for Java 5?

[joneshenry]

Geir Magnusson Jr. claims in the FAQ that "While the Java Community Process has allowed open source implementations of JSRs for a few years now, Java 5 is the first of the J2SE specs that we are able to do due to licensing reasons." Looking at the license files for the API specification of J2SE 5.0 versus J2SE v1.4.2, I'm not sure what the difference is. It seems to me that there is still a requirement that one implement every specified class, and only every specified class, and that one passes the TCK. Apache Harmony claims that they will be able to apply for access to the TCK under the exception for non-profits, but what if Sun doesn't see things their way? Without passing the TCK, a project doesn't seem to have any right to implement the API.

Before implementing the specification exactly and passing the TCK, there can never be an official release of Apache Harmony. I think this project is hopeless.

Incompatible

[theskullboy]

Couldn't, in the long run, this cause a compatibility issue? If a dev uses Harmony, and opens it up, but another dev uses the Sun implementation...wouldn't the API's be slightly different, causing a headache? Or will the API's be exactly the same (For further review, try running a mono c# app in .net...unless you use the common APIs between the two, there is no compatibility.)

I personally think it's a bad idea...Apache, watch out for the SUN!

Re:Incompatible

[Anonymous Coward]

unless you use the common APIs between the two, there is no compatibility

That's a rather large qualifier, especially considering the amount of overlap between the two.

Re:Incompatible

[anarxia]

That's why Sun does not allow you to call any implementation "Java" unless it passes their test suite. One of the major requirements is that you have to export the exact same api for java.* and javax.*

Java's biggest hole is in the embedded market.

[btarval]

You may well be right for PCs and Servers. But frankly, the current licensing scheme of Sun's Java is really getting in the way in the embedded space.

What typically happens is that some company has a neat idea for an embedded device. But they quickly find that the Java applications they want there won't fly because Java isn't supported on the hardware they were planning to use. They either have to fall back to a different CPU (which is usually more expensive), or pay a lot just to put Java on the CPU. Or go with C/C++ for their applications.

x86 and PPC are simply not the entire embedded world. There are many other superb (and cheaper) solutions out there, in this space. And no, Java support is far from prevalent on MIPS processors, despite what MIPS might try to claim. I know, as I've been there.

Please keep in mind that there are far more embedded CPUs around than there are PCs or Servers. So there is a clear need for Java, if it were available in this space. But it's not. gcc however, usually is, fortunately.

If Java were indeed Open Sourced, it just might be as popular as gcc is in the embedded space. Until then, people in the embedded space have far more flexibility by going with C/C++ than with Java.

Re:Java's biggest hole is in the embedded market.

[obi]

Yea, and to add insult to injury, even the PPC you mentioned doesn't have adequate support (no sun j2re) if you stray from MacOSX.

Write once run anywhere my ass.

Re:Java's biggest hole is in the embedded market.

[roadrunnerro]

You're mostly right, but projects like Kaffe aim to fill that gap too (ARM/XScale, MIPS...)

Re:Java's biggest hole is in the embedded market.

[mangu]

keep in mind that there are far more embedded CPUs around than there are PCs or Servers.

That's true, but there is one reason that's more important than licensing for Java not to be available for many processors. The most common embedded CPUs, some of which cost a dollar or so, are really limited in their capacity. There are C compilers for PICs, for instance, but there's no way they could run anything even remotely ressembling a Java environment.

Re:Java's biggest hole is in the embedded market.

[btarval]

I agree, but there are quite a number of CPUs which do have the horsepower to run Java. Enough so that the selection of a non-Intel CPU is feasible and reasonable. It is a shame to see people restricted here simply because they thought Java was originally a good choice, based upon their experiences with x86.

I suppose that does give a competitor who knows what it is doing a leg up. But still, the point remains that the lack of Openess with Java is hindering its adoption on a good many CPUs.

As far as the low-end goes, we're starting to see serious strides here. There are now $3 32-bit Microcontrollers appearing. Granted, the horsepower is only on par with the 1990 CPU's. But if Moore's law holds, we should see Java capable CPU's in this space within 10 years (assuming Java doesn't get horribly bloated).

Re:Java's biggest hole is in the embedded market.

[Anonymous Coward]

Well, they better hurry up, because .Net is rampart. So better to use the MS strategy of giving it away now, then clawing the 'next version' back later.

Re: Did you say Har Money?

[Anonymous Coward]

IBM connection? Perhaps there is the smell of money in the air?

Open source, but for a fee, I'll service it for you.

Thanks, Bill

[Doc Ruby]

Another preposterous propaganda for Microsoft apologists: MS monopoly abuse forces their impotent competition to open their source to survive.

That may not always be true...

[Kjella]

My point is essentially that it offers no advantage to the Apache development team to expend any effort in ensuring that their license is at all comptabale with the GPL because they will not foster anything from it.

...it is quite possible for several OSS projects to benefit from the same code, let's say a BSD licensed library if the GPL projects plays nice and releases their improvements of the BSD code under the BSD as well. But yes, there is the possibility that someone will make a GPL-only version of it, so you have no guarantees.

Kjella

Re:Possibly poor foresight.

[aliquis]

I don't know so much about the apache license, but anyway, not everyone wants GPL you know. I for sure don't. First because I can't look into someones code, "borrow" some implementation and use my code for whatever. Secondly if all computer software was GPLed how would any programmers earn their money? Donations? Code stuff for money because the functionality aren't there in the first place?

Why Apache Harmony is not needed

[Anonymous Coward]

Classpath permits commercial use via GPL special exception. As such, Harmony is useless project. Support Classpath, GCJ and Kaffe instead.

Re:Why Apache Harmony is not needed

[BohKnower]

I can't trust Kaffe, in its own site is said that Kaffe cannot be used as a replacement to Sun's JVM. To be well succeed harmony should catch up with JCP (Java Community Process) as soons as possible to have a J2SE 1.6 implementation at same time as Sun's.

Re:Possibly poor foresight.

[mcc]

No.

The GPL is the license the market has picked. If Apache willfully chooses to spurn that, that's their problem. And we, the consumers, are the ones who have to deal with the side effects of the problem Apache has created.

Of course they bless it!

[Anonymous Coward]

They know it will be years (if ever) that is comes about at any level of usability.

Re:Possibly poor foresight.

[Anonymous Coward]

Not the Java market.

Re:Possibly poor foresight.

[mcc]

The "Java market" doesn't so much exist at the moment, in the sense of a viable choice of JVMs for production use. This is, one would expect, one of the things this project would hope to fix. One would also expect this would be somewhat difficult if you select a license which precludes the bulk of important open source code out there today from linking against you. This is a problem. As a platform, one might even say a piece of middleware given the JVM's nature as a negotiator between software and operating system, one would expect the JVM to select a license which makes it as flexible as possible despite any political considerations.

in soviet russia

[Anonymous Coward]

they host duke nukem 3d on a beowulf cluster powered by open source java.

Re:Possibly poor foresight.

[KarmaMB84]

If the OSS community thinks people should have freedom to do anything with their software, shouldn't that freedom extend to a choice of licenses? If it's open source, it's open source. It will still satisfy the goals of the end-user having control and access to the source. Whining about licenses usually comes from people who want to lift code and put it in othe projects and never contribute back to the original source. For proof, look at any announcement about an OS going open source and see people yell and scream for it to be GPL so they can lift code for Linux (not to contribute back ofc!).

Re:Please Add Decimal Math (fixed-point) as a buil

[KarmaMB84]

And just like the last language I heard of having such a beast, nearly nobody will use it thus making the effort implementing it a waste.

The coming copyright violation lawsuit

[joneshenry]

The specs are open but the license file for the API specification specifically grants only the right to implement the entirety of it, and only if the implementation also passes the TCK. There is no other right to use the API specification other than for "internal use" to be able to write programs that are not re-implementations. The effort to achieve total compliance is simply beyond what anyone can achieve without the ability to reach intermediate steps. But those intermediate steps violate the license of the API specification by not being complete.

What should in my opinion be alarming is that Sun has specifically mentioned a Residual Knowledge right for its otherwise "read only" JCK that is not mentioned at all in the license file for the API specification. That means that Sun considers one to have no Residual Knowledge right after having read the API specification. Anyone who reads the API specification is tainted unless the implementation is 100% complete and passes the TCK, which isn't going to happen soon, or ever. This is the same clever strategy that Sun has used of spreading the source code to taint as many developers as possible. And while GNU Classpath and Apache Harmony can probably succeed in documenting that developers aren't tainted from having viewed the source code, there is no way they will be able to argue that all of the developers are also untainted with respect to the having read the API specification.

I think the ground is being laid for a straightforward copyright violation lawsuit to defend the API specification should a release ever be made of either GNU Classpath or Apache Harmony. There is no way either project will achieve complete compliance, and there's only so long they can claim they are "pre-released" products, especially if they are being widely installed and used as dependencies in open source distributions.

You might not think Sun can win such a lawsuit, but it doesn't really matter now does it from an individual developer's perspective if one doesn't have the legal resources to fight back.

Re:The coming copyright violation lawsuit

[Anonymous Coward]

I don't see how an API is copyrightable. The old copyright exception of allowing copyright violations when they are inherently required for compatibility kind of requires it. Otherwise Compaq couldn't have implemented the IBM bios and the world would be a very different place.

Also, I think you are really misreading the license. A residual knowledge clause like this one basically is waiving the argument that you are "tainted" for the purposes of implementing other software.

Re:Possibly poor foresight.

[soulhuntre]

The GPL is the license the market has picked.

No, Apache is. Many many more companies run Linux as a free way to run Apache than use Apache only because the chose Linux.

Re:Possibly poor foresight.

[mcc]

Depends on whether people want to be a member of the open source community or outside of it.

BS

[Black-Man]

I *daily* move Java source from Windows, Solaris, and Linux. I have not encountered anything in Java that breaks. A third-party lib like MQ... maybe. But then again, this is the REAL world and MQ doesn't play with LISP. And neither do I since I graduated from college.

Re:This is great news...

[teknikl]

I know I did that with v6, Can't be arsed to do it again for v7. Honestly I know my post got modded flamebait but it was meant to be funny. Java started out as a decent tool - but the implementation of the VM is so off (result of lawsuits?) that it takes 8x longer than a flash anim to load. That seems wrong to me java should be portable robust AND lightweight.

Re:Possibly poor foresight.

[C_Kode]

That doesn't make the GPL bad. I personally like the GPL but I find the BSD license just as valuable. Don't hate the GPL (not saying you do) It brings a lot to the table that you wouldn't have otherwise. Look what advances the GPLing of the Linksys code has done and thats but only but a fraction. Hello IBM.

I love FreeBSD. I'm unexperienced at the other BSDs (used them for a short period of time though) but Linux has a greatness also. Tell me what you need, and I will tell you what you need. BSD and Linux (or better noted, two difference licenses) can handle your problems. Both have answers, but advance in different ways.

Why? Different license, not GPL

[UnapprovedThought]

This attempt is placed under the apache license, which is similar to the BSD license. I'm not a legal expert, but basically I gather that it means that the code can be reused and enhanced by commercial organizations without them having to offer the source code changes up for redistribution.

Thus, if you contribute under this license, MS (for example) can run off and add your code to .NET, distribute it as a binary, sell it, and not have to show you the changes (would it lead to a future patent that would come back to bite you?). Under the GPL, they would have to redistribute source, which would allow you to learn what, if anything, they changed. Requiring the source to be open would allow you adapt the changes back to an open standard if necessary.

Maybe releasing it under this license is the only way to get enough cooperation from Sun or other patent owners to get such a project to work?

And a related question -- what were Ballmer and McNealy smiling so much about? Well, I'll leave that exercise to the reader.

Re:HEY ASSHOLE, CUT ZONK SOME FUCKING SLACK!

[Anonymous Coward]

The worst part is, he doesn't stop sodomizing you when you're unconscious. :-(

Ed Seminoe, former Andover.Net intern and sodomy victim.

Is anyone else reading a fork implication here?

[shaitand]

Did anyone else read this as, "java gods declare sun evil, fork and establish new and open authority to replace them?"

Something about the overall tone seemed to imply that they weren't just writing an implementation, but intended it to supercede Sun's closed implementation.

Sounds good really.

Re:Getting rudiculous

[Anonymous Coward]

Zonk, you are an embarassement, an icon of shame for this website.

lick my balls

-Zonk

Misguided

[Anonymous Coward]

They might adopt the GNU Classpath class library.

That'll happen - not. You expect them to chase down the 100 major contributers to Classpath and have them sign over all their source code to the new non-GPL license? You're living in a Java dreamworld! But I like your positive attitude.

Re:Misguided

[Anonymous Coward]

That'll happen - not. You expect them to chase down the 100 major contributers to Classpath and have them sign over all their source code to the new non-GPL license?

All code in GNU Classpath is already signed over to a single source: The FSF.

And the current license isn't the GPL.

Java Compliance Testing

[narl]

The only restriction is that you can't call it Java unless it meets the spec (and proving that it meets the spec is, quite understandably, nontrivial because Java is a large, complex language).

I probably have this wrong, but I thought the problem was that while the Java Spec is open, the Java Compliance testing is not, so to get your implementation tested is a pretty expensive barrier.

And do you have to get your implementation tested on every architecture you want to support? That would really hurt.

Why to use Java

[narl]

Java is the number one development environment for business applications. Bar none. Bahahaha. Number one in what? Being slow? Being broken? Being inconsistent? Being verbose? Being a nightmare for sysadmins to manage?

From what I can tell, the reason Java seems becoming the number one business applications (bleah), is that the Java language structure does a pretty good job of straight-jacketting you into writing somewhat maintainable code.

From the PHBs' point of view, more maintainable code is more valueable than more efficent/faster developed code because it means they can treat the developers as interchangeable units they can add or remove on demand. Filling a Java opening is easy, filling a Lisp opening would be a nightmare.

Time for a new Section on Slashdot

[SilicaiMan]

... named 'Dupe'.

Re:Possibly poor foresight.

[mangu]

more companies run Linux as a free way to run Apache than use Apache only because the chose Linux.

No. If you choose Linux, you are stuck with Apache, but if you choose Apache you don't need to use Linux. There are other free OSes that run Apache, but there's no easy way to run Linux with the second most popular web server.

wasting time in java

[ghostzerg]

dont needed. python is already open-source.

Re:Possibly poor foresight.

[aliquis]

As a user I like the GPL, but I still think it could/do lead to an inflation in the value of code/programming. Say for example some developer helps with firefox, because he/she can and think it's fun, the problem is the user doesn't earn any money, but companies like RedHat do.
Some day the said developer might want to earn their living on programming, the problem then is that if he/she for instance starts to code his own browser, well, the chances are much lower anyone will buy it.

Also of course it's great noone can steal others free work and GPL code, modify it and earn money of it without contributing, but the same thing will also lead to some people/companies not improving or using the software. Which gives less choices.

I have no idea how one could solve it thought. Maybe some open-source shareware license where you pay a royality to the one you borrow code from.

Re:Possibly poor foresight.

[aliquis]

For example I guess the people who have worked with Solaris sees the evolution of Linux on the server market as something which isn't only good.

yes

[namekuseijin]

"can you run GPLed Java software on the Sun JVM?"

why not? the JVM is an emulator for an idealized computer. just as i can run Linux on any computer without regards to its firmware being proprietary, i can also run GPLd java code in Sun's JVM.

and there's also Classpath exception when binding code. Besides, you can run it in Sun's JVM using classpath code as the libraries, so as to be completely free and also notice the output of GCJ can be given any license you wish, just like GCC's output.

Re:Possibly poor foresight.

[Anonymous Coward]

That's funny when you consider that the reference development platform for the apache team is FreeBSD. Same for bind and sendmail, btw.

I read this on java.sun.com

[tod_miller]

Sun are happy about this for drakes coffee cakes!

To be honest, I think the mono project has a little more credibility in terms of 'filling a hole' but, and this is a JLO style caveat, this project has greater significance in the linux development community.

NOW before you start getting your undergarments of choice in a complex multi-dimensional distortion, this is the VM. You know, the thing that ust be 100% compatible blah blah.

Now, the libraries on top of the VM, are SCDL..SCLC... whatever thier acronym is... I do not think a GPL, LGPL or anything would be of benefit.

If took IBM to make a pretty neat VM, I think apache may have a token gesture, but I do not see anything usable (by usable I do not mean Hello World) within 24 months, or anything desirable at all, ever. Since thier catch up time, we will have a newer Java, with all kinda of neat features (TM) in terms of performance, the way I see it the next major VM release isn't going to be too cold on the heels of Java 1.5.0_02 / 5.0.

I am still using 1.3.1_02 for one project, and skipped 1.4.* entirely for this project and I am just hitting 1.5.0_02 for it. But this goes to show how utterly l33t Java was even at 1.2/3

And now, because this is slashdot:

FreeBSD, no women, soviet russia, old korean people, all your base, goatse, dupe, whore, puke-paille.

Troll!

[Anonymous Coward]

>>they still push their proprietary operating system
Solaris 10 is Open Source with and OSI approved mozilla-type license
>> (the current version of which is 30% vaporware, >> btw)
No, ZFS and Janus (Linux app compatiblity) are sheduled for release later this year. They are in an advanced stage of development, so no - they are not vaporware.
The rest of the features Containers, DTrace etc. were released months ago.
>>They are also losing server market share
They are the leading seller of AMD Opteron servers. Expect to see fireworks when they release their new servers towards the end of the year.

Re:Troll!

[iggymanz]

sorry, but that janus/zfs stuff has been pushed to 2006 now. Solaris 10 open source? hahahaha, they've only released dtrace thus far, can't boot my machine on that. Sorry you drank so much of the koolaid, you didn't notice it's got no flavor and no sugar.

Re:Getting rudiculous

[zaxios]

lick my balls

Okay, Zonky Wonky, just turn off that silly computer and come back to bed.

Re:Possibly poor foresight.

[Anonymous Coward]

The Java Market doesn't exist? Sounds like propaganda from the Free Software Animal Farm.

I think you have an argument for the JVM itself, which might need to integrate with (L)GPL C libraries. But for "userland" Java libraries, the Apache licences is much more established -- compatibility with things like Tomcat/Struts/Geromimo is important.

Also, I will note that the Apache Licence has not exactly been a huge problem for Open Source web development platforms.

Re:Possibly poor foresight.

[m50d]

No, the Apache license responsible for the incompatibility because it was written *after* the GPL. You can't seriously expect the GPL to be written to be compatible with future licenses.

They're putting Sun under pressure...

[vhogemann]

Apache is a key member of the Java community, I guess that Harmony project will put some pressure on Sun to opensource more of their code.

See.. if I can run Geronimo, Tomcat, Struts, Tiles, Velocity and many other projects under Harmony, why should I install Sun's VM? Harmony could well be crafted as a modular java implementation, I don't really need all that swing classes on my webserver, and that is a major selling point.

I could guess that once Harmony get a functional core, Sun would start donating code under the Apache licence... witch is more friendly to their "corporate ways" them than GPL.

Let's wait and see...

Forgive me, but...

[John Nowak]

If you need to write applications that run on platforms not supported by Java, er, how about not using Java? There are certainly many other solutions out there that are quite lovely. Sun's Java does a good job on the important platforms. How many people run PPC Linux? Not so many. I do, but not having the latest, greatest Java VM is hardly killing me. My point is that the problem is not so big that it warrants wasting all this time, especially since other people are already working on solving the problem. Compatibility is always going to be an issue with with the OS VM versus Sun's VM, so why not just do something entirely new altogether? You can't seriously think that Java is the best we can achieve. Beating Sun at their own game is impossible. Instead, let's create something new if it is really necessary, which I'm not convinced it is.

Does this work as an alternative JDK/JRE?

[Jesus_666]

Can Jikes be used as a complete JDK/JRE? I'm using an iBook for most of my development stuff (because I can use it to code during boring lectures) and I don't quite feel like paying Apple 120 bucks for the Java 1.5 SDK. Could I replace the Apple JDK with Jikes and expect it to compile/run most 1.5 stuff?

I don't really care about performance as I rarely need to run any Java apps besides the exercise stuff I write.

Re:Does this work as an alternative JDK/JRE?

[bth]

I do not know the current status, but when it left IBM Research, it did not have the full Java libraries as part of the open source package (one of the reasons it was called an RVM).

Re:Does this work as an alternative JDK/JRE?

[Anonymous Coward]

Couldn't you use Eclipse's compiler? With 3.1 it'll be 1.5 compliant and all you need is a JRE

Gluecode not connected

[mparaz]

Geir Magnusson Jr. replies here:

Apache Harmony has nothing to do with Gluecode. Gluecode focuses on Java application servers. It has nothing to do with J2SE implementations, and has no interest in J2SE implementations.

a sad state of affairs

[willCode4Beer.com]

I think we've come to a really sad state when a bunch of guys working for free part time as a hobby need a bunch of lawyers to tell them how to go about it.

Old MS JAVA SDK

[Anonymous Coward]

Does anyone know where to download the old MS JAVA SDK? It's needed for some old non-standard code...

Re:Please Add Decimal Math (fixed-point) as a buil

[Anonymous Coward]

Have you heard of Visual Basic? It has a fixed point datatype.

Thank you!

[Safety Cap]

"A group of 12 Apache developers have put together a proposal called Harmony. The proposal appeared as a simple project call last Friday on an Apache incubator mailing list. It would make this new, built-from-the-ground-up version of Java available under the Apache 2.0 free software license."

This article advocates a

( ) technical ( ) legislative ( ) market-based (x) vigilante

approach to open-sourcing Java. Your idea will not work. Here is why it won't work.
(One or more of the following may apply to your particular idea, and it may
have other flaws.)

( ) 7337 h4xx0r5 only work on k3wl k0d3
(x) Projects left on "Development Status: 1 - Planning" do not attract any serious developers
( ) Only Microsoft can produce a product with only FUD/hype and no actual code.
( ) It is defenseless against existing products
( ) It will be stable for two weeks and then fork too many times
(x) Users of mainstream hosting providers will not put up with it
( ) Microsoft will not put up with it
( ) Sun will eventually not put up with it
(x) Standards committee will bicker endlessly without writing any code
( ) Requires too much cooperation from too many people
( ) Requires immediate total cooperation from everybody at once
(x) Many Java developers cannot afford to learn the JDK-du-day
( ) Poor documentation, the hallmark of FOSS projects, will kill it
(x) Lack of any existing code will discourage participation

Specifically, your plan fails to account for

( ) Programmer apathy
( ) Lack of centrally controlling authority for codebase
( ) Existing versions work 'well enough'
(x) A committee without code is like the US Congress: lotsa talk, no action, then gets everything wrong
( ) Asshats
( ) Unpopularity of weird new ideas
( ) Public reluctance to accept new compilers
( ) Huge existing software investment in jalopy/javac
(x) Susceptibility of untested new codes to attack
( ) Willingness of users to install new jdk
( ) Armies of worm riddled broadband-connected Windows boxes
( ) .NET
(x) Technically illiterate PHBs
( ) Extreme stupidity on the part of the lUsers

and the following philosophical objections may also apply:

(x) Ideas similar to yours are easy to come up with, yet none have ever been shown practical
(x) Any scheme sold soley on the fact that it is FOSS is unacceptable
( ) Native code running on one platform should run on any other platform
( ) Why should we have to trust you and your servers?
( ) Incompatiblity with standard JREs or Sun's licenses
( ) Feel-good measures do nothing to solve the problem

Furthermore, this is what I think about you:

(x) Sorry dude, but I don't think it would work.
( ) This is a stupid idea, and you're a stupid person for suggesting it.

Trekkie Goons Go Home

[sabat]

As Spock once said in a Star Trek novel "A difference that makes no difference is no difference".

As William Shatner once said, "Get a life! Will you people!?"

Seeing implementation

[a2800276]

If you're just curious, you can see the a java implementation now. Apart from all the open source versions, each Sun JDK release comes with a file called src.jar that contains the entire source of the J2SE API classes. The source to the JVM, compiler, etc. are available to look at after agreeing to a click through license here: http://tinyurl.com/e4dho

Re:Possibly poor foresight.

[Anonymous Coward]

And of course you had to choose the two most bloated web servers in existance.

I'm currently running thttpd with a 6-line configuration file that sets up exactly the same things that I would need months to figure out how to do in a several pages long apache config file.

Re:Not Embarassing (OT)

[Anonymous Coward]

he bought it on ebay