Sun Microsystems has expressed “serious doubts” about the usefulness of the latest Apache Foundation project to create an open source implementation of the Java 2 Standard Edition (J2SE). In an interview with vnunet.com, James Gosling, Java creator and Sun vice president in charge of the programming language, explained that he did not understand why the open source consortium was undertaking the project.
Sun lashes out at open source J2SE
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
“he did not understand why the open source consortium was undertaking the project.”
Because they can?
I don’t think Sun in general gets O.S.S..
Today they don’t like opensource, tomorrow they will like it, in yet another day they will think whatever way that necessary get them another headline proving that Sun still exists.
All software is turning open-source. Shareholders will have to look somewhere else to invest their bucks.
I agree with the OSS movement to make an open source J2SE, I’m a former user of the Java language because the lack of a Java interpreter on most distributions made it a pain to redistribute programs, so much so that it was virtually useless. Sure there is a version of J2SE out there for Linux already but because its not OSS many distributions are shy about including it, and I heard from someone trying to get Java into a distribution’s default packages that one of the requirements is that you not include anyone else’s Java compilers or interpreters.
If an OSS version of J2SE isn’t good it won’t survive, if its not compatible with Sun’s java most people won’t consider it good and it won’t survive, therefore I don’t see any compatibility issues arising except for a slight lag in version numbers for the non-Sun implementations. The fact is that the majority of people in Linux want an open source java implementation so they can include it into their distribution without having to worry about anything, this is a perfectly reasonable expectiation if they are willing to write it themselves and at the rate thing are going it will be out far sooner than Sun’s “open source” version of Java.
I’m sure people will dissagree and that’s their right; however, most of the people here tend to engage in the tradition of arguing about something they know nothing about. RedHat is a distributor who wants open source Java, Novell decided to start the Mono project instead of waiting, and several other distributions are fed up with how Sun Microsystems is making it difficult to include the language. I’ve done some volunteer work for Ark Linux and I’ve worked at lot with Java in the past and I can tell you that all those distributors can’t be wrong about wanting to have a version of Java which they are free to redistribute without having to make any unreasonable commitments.
The OSS community has been talking about making an OSS version of J2SE for a long time; projects such as kaffee, jikes and a few others have already attempted to re-implement J2SE for a long time and still Sun Microsystems didn’t release a version of Java that was licensed favourably to the OSS community. How can they now complain about a situation which they themselves are responsible for?
I hate to sound like a troll but Sun Microsystems is ignorant beyond imagination at times.
is because of comments like this.
OSS people asked for Java to be opened Sourced, What does Sun do create a new license and Open Solaris.
IBM grants OSS developers access to a large number of Patents What does Sun do? releases their patents to their License and says GPL doesn’t make up all OSS programs.
Sun Pushes Open office to use java tightly, forcing one to install Java to use Open Office. Then complains when OSS users start making their own version of the supposedly ‘Open” Java framework.
The title says “Sun lashes out at open source J2SE.” My interpretation of the article is they think it’s pointless and don’t get why they’re doing it…but I don’t think they’re “lashing out.”
What I dont like about the harmony project is that most of the motivation seems religious/political. A lot of the original discussion was just repeats of “it’s not open source, sun won’t make it open source” over and over again. Reading the article it says that with an open source implementation they can port it to more systems than the sun vm currently supports, that’s the first thing about it i’ve heard that’s actually made me think the project isn’t completely uselss. Maybe they should’ve advertised more points like that from the beginning.
So not only does OSNews just link to someone else’s story as if it were their own they also copy the arguably incorrect headline.
1) It wasn’t Sun in that interview. It was James Gosling
2) He didn’t “lash out”. He explained that he didn’t understand why the Apache was doing what it was doing. The man is entitled to voice his opinion.
Every day Sun either loves or loathes OSS. It’s crazy.
I’m also very interested in an open source J2SE, being able to just have it pre-installed in my distro of choice instead of going through all the crap on the Sun website makes it worth it in itself. Maybe it’ll even force Sun to loosen up on their licensing.
It seems that we’ve got some in the open source camp working on an open source implementation of Java, and others are doing the same thing with .NET. Seems to me that the OSS community would be better off telling both Sun and MS to go take a flying leap and come up with something of their own – something even better.
>>>IBM grants OSS developers access to a large number of Patents What does Sun do? releases their patents to their License and says GPL doesn’t make up all OSS programs.
You people are too kind to IBM and too judgmental to SUN. The reality is always somewhere in between.
For the record, SUN actually spent cash — more than anybody in the whole world — on opensource stuff. Buying staroffice for $74 million and open-sourcing it.
But the problem is that the more money SUN spent on open source stuff — the more confused their overall open source business plan becomes.
IBM spent very little money on open source stuff — certainly not the 1 billion dollars that they advertised. A lot of that money goes into getting more linux functionalities into AIX. Even IBM’s eclipse “donation” was advertised as only $40 million — less than what SUN spent on buying staroffice.
IBM made “donations” of their codes vs. SUN paid actual cash to buy staroffice and set it free. The irony — SUN paid actual cash but still get trashed by the open source community for this and that.
What IBM has — is an intelligent business strategy on linux and open source. IBM can grant you the 500 patents that they have, but IBM is afraid of the patents that they don’t have. So IBM doesn’t make their own distributions because of patent concerns. IBM wants to sell you their powerpc-based hardware for embedded markets, but they also don’t want to embed linux into their own commercial products. In the end, IBM wants to sell IT services.
Sun suffers from this awful problem. They had a brilliant group of technical people put Java together. Top notch engineers and developers. Aside from that though, they seem to have management that is clueless.
“[Apache] says a lot of words about why they want to do it. Exactly why is it critical to have a delta between our source licence and the source licence that they think is appropriate?” he said.
It’s like Sun management doesn’t even realize *why* folks like to work on free software.
“We hardly have the energy to work on our [J2SE implementation]. We’ll be glad to get co-operative and helpful, but there is only so much energy that is free and donatable,” Gosling told vnunet.com.
Ha! Heh! When you make it free, folks will *volunteer* to help work on it.
The Apache folks aren’t even interested in a fork:
In response to Gosling’s remarks, Geir Magnusson, an independent software developer with the Foundation, told vnunet.com that Apache does not aim to fork Java.
An open source J2SE implementation could allow the software to spread to new devices, according to Magnusson, who pointed out that Sun’s J2SE only supports Solaris, Linux and Windows.
And Sun *still* wants nothing to do with free software. Yow.
BTW, notice how nowhere in the article GCJ is mentioned. Why is that? Seems like an attempt at misdirection to me. Sun *has* to be aware of what’s happening, no?
Well, the whole point is rapidly becoming irrelevant anyway. GCC/GCJ 4 just came out and Java support is better than ever. Eclipse, Ant, and Tomcat will soon run right OOTB (that is, without any patches or futzing around) if not already.
Very soon, it will be just as easy to install and to get full Java support from GNU (GCJ) as it is from Sun. Then folks will begin to wonder what Sun is still doing here.
Don’t forget Azureus! It will run on FreeJava (Kaffe/GCJ) with a 3 line patch.
Who needs Java when we have D anyways?
http://www.digitalmars.com/d/
Spreading to new platforms is one of the things I’d like to see. I’d love to be able to use my Java skills to write palm applications but I don’t even know where to get started because I’ve found nothing to do with the javax.microedition.* packages in the J2SE SDK and I’m not sure where else to look. Perhaps I can find some instructions in the online Java books but truth be told I’d rather just stick with something familiar such as Swing even if it can’t be fully implemented for the palm devices.
I’m hoping that with an OSS implemetation of Java one of the things I will see is more documentation for writing Palm apps in Java and implementations that are more complatible with desktop java such as using even a partial Swing instead of a whole new toolkit.
Sun suffers from this awful problem. They had a brilliant group of technical people put Java together. Top notch engineers and developers. Aside from that though, they seem to have management that is clueless.
Ummm…you do realize that Gosling (the man from the interview) is in fact the ringleader and founder of the “brilliant group of technical people put Java together” right?
I’m in a training class this week for java, while everyone is using the supplied M$ pc’s with Sun’s Java, I’m using my rawhide laptop
Everything I’ve done so far as worked fine.
$ java -version
java version “1.4.2”
gij (GNU libgcj) version 4.0.0 20050519 (Red Hat 4.0.0-8)
Copyright (C) 2005 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
javac -version
Eclipse Java Compiler 0.548, pre-3.1.0 milestone-6, Copyright IBM Corp 2000, 2005. All rights reserved.
You got to love OSS/FSF
One of the main problems is that Sun isn’t responsive to the community. PowerPC linux people have been begging for an official PPC version of J2SE from Sun for years. Right now, you have to go install an IBM version put together for POWER systems.
Another issue I recently ran into involves SMP. I recently put another Opteron on my mobo. Everything else works great, but now JAVA apps abort on startup. After looking around the net, it seems this is common problem with JAVA and SMP, and has been for years. I had to install Blackdown’s 1.4 JAVA to be able to run JAVA apps with SMP. Thankfully, Blackdown is working on a 1.5 version as Sun certainly doesn’t seem to care about this.
As mentioned above the title given is extremely misleading, the opinion of one employee of Sun is not indicative of the companies stance. For this one voice against Harmony I can name at least 2 people from Sun who have expressed enthusiasm about project Harmony.
I disagree with a few statements mad by Gosling here, an open source J2SE has a number of benefits over Sun’s J2SE:
– It can be distributed/packaged with F/OSS only distributions such as Debian.
– The community can take the implementation to platforms and markets that may not be financially viable or beneficial to Sun.
– It takes away at least some of the vocal “open source your java” crowds arguments.
What James’ says about the enerprise market however I agree with, a majority of these companies want a quality product with quality support offerings, something they are already getting.
Really I think that if Sun is going to get behind any open J2SE implementation it should be this one. Apache are known for great quality software and their license is very reasonable. I am sure the GNU guys could create something just as good however the GNU license just doesn’t do it for me.
Let me spell it out for ya, Sun:
LICENCE
apprently someone has to spell it for me
License
Either Harmony will produce a totally free Java JVM and class library in five years (they don’t have to DESIGN anything, just code to the existing spec!) or they will force Sun to totally OSS Java.
Either way, we win. We get a totally free OSS Java. If Sun has a complaint, who cares? Let it play out.
This is the same complaint Linus made about Tridge over the BitKeeper nonsense – that OSS undermines developers by reverse- or re-engineering commercial stuff. The argument made no sense when Linus used it and it makes no sense here.
The bottom line is: this is the way economics works – it’s all a case of “What have you done for me lately?” Don’t try to rest on your laurels because your laurels are being copied and re-engineered and sold by somebody else. The point is to keep moving forward and rely on innovation and speed (and if necessary, technological and market inefficiencies) to get your monopoly profit.
It’s called “Darwinian evolution” and if you can’t keep up, you’re a dinosaur. Or as they say, “If you’re not part of the steamroller, you’re part of the road.” Or as Linus quipped about Gates’s “The Road Ahead”, “Anybody standing in the road looks like roadkill to me.”
Because they can?
Harmony is complete vaporware at this point. The odds of coming up with a complete J2SE5 environment (which is their goal) is slim to none….unless someone like IBM gets behind it.
But you can always count on open source duplication of effort..over and over and over again.
So 3 years down the road in the small chance that they are successful, we’ll see Sun’s implementation 3 generations ahead, much better performing, more complete, and everybody except zealots will still be using Sun’s version.
If they’re going to waste all this time, why don’t they shoot for something better than Java. Hell, fork Mono, extend the CLR, and come up with a truly generic runtime. At least the Mono people are implementing something that doesn’t exist on Unix to begin with.
Well, the whole point is rapidly becoming irrelevant anyway. GCC/GCJ 4 just came out and Java support is better than ever. Eclipse, Ant, and Tomcat will soon run right OOTB (that is, without any patches or futzing around) if not already.
Oh yeah, RedHat has fulltime people working to make sure that Eclipse works OOTB, and then when someone uses the real deal, tries to compile it with GCJ, and it inevitably fails to work, you’ll have people not caring about GCJ or thinking its buggy crap.
I wonder how many million man-years have been wasted because things aren’t “open” enough.
No wonder the open source desktop fails to make any headway with so much wasted developer time.
The only valid argument I found from Apache Foundation’s side is that their J2SE implementation will support more platforms than supported by current implementations. But does that justify an Open Source J2SE implementation? What purpose does an Open Source J2SE implementation serve? -apart from letting programmers like us fiddle with the source
Gosling’s apprehension about J2SE forking also seem unfounded since Apache is willing to make Harmony undergo Sun testing for certification.
Does an Open Source J2SE implementation imply that Linux distributions will integrate it? If yes, the project seems justified.
Lastly, I agree with others that the title was indeed misleading, Gosling’s reaction was more of indifference rather than lashing out.
“No wonder the open source desktop fails to make any headway with so much wasted developer time”
– Now…lets just think about this for a bit. If they hadn’t taken this approach it wouldn’t really be an open source desktop now would it?
Gosling-i’m just a fat man with a beard who wrote a language to turn on microwaves in the early 90s. I don’t know why the most inluential open source community would want to rewrite a programming language that i wrote decades ago…did i miss something? is there a new microwave coming out?
Yeah, let’s think about this. No “approach” has been taken. Harmony is complete vaporware. No code had been written. So instead of improving gtk+, Gnome, you’ve got RedHat developers duplicating Suns efforts. And now you’ll have who knows how many other developers duplicating Sun efforts instead of doing something novel, implementing something that doesn’t exist on Unix, improving Apache, or whatever that was invented in the open source world. There’s already free Java for Linux. So once again, because we have some people that think source code is a religion, we have who knows how many hundreds of thousands of man hours that will be wasted. They’ll be 3 generations behind by the time this is even usable.
The only saving grace for this project is that Apache has a lot more credibility with companies like IBM than GNU has, so really the only hope is that IBM gets involved. GNU will have to change the license anyway to play with Apache.
It’s a shame Microsoft wasted so many man-years reimplementing Java because it wasn’t “open enough.”
Ted, grow up and actually read the article.
Tech news sites are mostly rubbish. Vunet interviews this guy, who expresses an opinion that isn’t extreme or bizarre or anything like that, and then they come up with an ad-revenue grabbing title and paraphrasing that bypasses sound journalism entirely. The article even manages to contradict itself: “Sun will not contribute to the project, Gosling said….” VS. “‘… We’ll be glad to get co-operative and helpful, but there is only so much energy that is free and donatable,’ Gosling told vnunet.com.”
Other people at Sun in blogs, etc., have expressed some enthusiasm for Harmony, so saying that Sun is lashing out is just nonsense. If anything, certified Java implementations shipping on even Debian or OpenBSD is _a_good_thing_. The key word is ‘certified’ meaning it will work and not be some half-ass project that never quite implements any API fully.
No language with an open IMPLEMENTATION has ever had the issues Java has. See any perl, ruby or python coders bitching about version X on platform Y not behaving? Nope. This versionitis strictly afflicts the closed runtimes. Open specs are of limited value because you end up with N implementations that vary ever so slightly – just enough to break YOUR code, because a spec is just a high level description with some tests.
If/when the open J2SE comes about, hopefully it will be the DE FACTO implementation. If not, it will in fact be a wasted effort. Really, perl, python etc are successes in this regard because no competing runtime was ever implemented…the first implementation WAS the de facto implementation.
>> Harmony is complete vaporware at this point. The odds of coming up with a complete J2SE5 environment (which is their goal) is slim to none
wrong, chances are 90% plus if they get the correct parties to relicense. If Classpath relicenses they are most of the way there. A runtime can be spit out in six months, or Parrot etc could be used, or one of the already-built VMs.
my impression is they have no intention of going from square 1.
I agree with you, there is an immense amount of duplication. One thing we have to keep in mind is this:
The “open source community” is a very loose term. We are all in it for different reasons and at different levels, it is innevitable that there will be lively discussion and disagreements.
So licenses are very important, particularly in regards to commercial use and “freedoms”, I understand that intellectual property can be an overwhelming issue but sticking our heads in the sand is not going to make the problem disappear.
The best thing that any level-headed person can do is take the middle road, understand the perspectives of all the parties involved and formulate an opinion from there.
Again, alot of open source developers are gonna waste their time by reinventing the wheel.
Anyway, that’s the only thing most open source people can do. Don’t get me wrong but it’s the sad truth. They are like music bands making cover albums instead of writing their own music. Sure they are good developers but they are not creative at all. Unfortunately, I think that creativity is the most important thing after all.
Most of the time, these guys take someone else idea and make a “cutted down” FREE alternative. I don’t get their philosophy. It’s not that I’m against open source, I just don’t get why they just duplicate stuff to give it out for free. Why not make something new?
The license file that accompanies the API specification lays out very clear terms for the only allowed usage: (1) Using it to develop programs that do not re-implement any part covered by the API (2) re-implementing the API and only what is in the API for classes that fall under the API namespace (3) having the re-implementation pass all conformance tests.
In other words, intermediate stage re-implementations that only achieve part of the API are disallowed. Think of how difficult it would have been for Linux to be released if it had had to meet every single one of the Open Group’s Unix specifications at release 1.0. And this is not a trademark violation, it is a straightforward copyright violation of the API specification’s license unless one can somehow prove one’s developers are untainted by having ever read it. Good luck on proving that.
Note that in the JCK “read-only” license at https://jck.dev.java.net/jck-read-only-license.txt Sun explicitly grants a “Residual Knowledge” right for reading JCK source. No such “Residual Knowledge” right is granted in the API specification license. In my opinion, Sun is laying the groundwork for claiming anyone who has read the API specification is tainted from re-implementing unless the license terms are met, just like anyone who has read Sun source code is acknowledged, even by OSS projects, as being forever tainted from contributing source code.
An OSS re-implementation is in my opinion in greater danger of being sued out of oblivion than Mono.
My memory is that RMS created the GPL as a reaction to Gosling forking the then emacs code into a proprietary product for which he would not share sources. There is a considerable difference between the generosity afforded by the GPL and the business model milking of Sun’s efforts in OSS. The GPL frees you from your vendor but Sun’s licenses tie you to them. The problem with Gosling at this point is that he doesn’t realize that Harmony presents the same dillema to Sun that x.org did to xfree. Lets see if Sun has learned the lesson that xfree did not: “if you can’t beat them, join them”.
“Man-hours” sounds like a philosophical discussion to me, who gets to determine whether a person is wasting them or not?
In my opinion your man hours are your own and the same applies for everyone, no one can own another man’s man-hours or tell them how to spend that time. Whether a person decides to spend that time helping themself, helping others or simply doing nothing the time is only wasted if the person who’s time it was decides.
Even in a job you are still the owner of your own man-hours, the reason your employer gets to tell you what to do during that time is because you have agreed to let him do so and you can always change your mind later as long as you don’t mind losing your job in the process.
Lots of people write programs for fun, I’m one of them and no matter what my program is, even if its just another version of the “hello world” program, no one can make me believe its a waste of time. The same thing goes for the people behind the Harmony project, if they believe its a good use of their time then it doesn’t matter what any one else says, the only opinions that are going to change project Harmony in any way are the opinions of its developers.
“I understand why they would like it to be different. From our point of view that would actually be more destructive than helpful. It boils down to forking: they believe that the ability to fork is an absolutely critical right.”
Heh, right to forking is absolutely critical to ANY open source software package. It may either cause harm or become panacea to many problems (gcc is a good example of a ‘good’ fork). If Gosling does not get it, he doesn’t get OSS as a whole. Personally, I don’t believe that Sun (or Gosling in particular) does not understand this.
IMO Sun has done (and still does) enormous contributions to Open Source, but loses all credits and confidence due to uncertain comments/decisions about OSS. I believe they have a problem switching their corporate culture (say ‘DNA’?) from selling packaged software/hardware (threatened by OSS and commodization) to services.
I don’t think that Harmony is a good project to undertake.
If they want to do something for Java thay can fix the millions bug that are unfixed.
http://www.sunsource.net/
Yeah right, Sun doesn’t get OSS.
http://weblogs.java.net/blog/kgh/archive/2005/05/thoughts_on_the_1….
http://weblogs.java.net/blog/calvinaustin/archive/2005/05/harmony_f…
Yeah right, Sun does not want Harmony.
Jeez guys, stop downing a whole company just for the few negatives things you hear about.
“If we could get the enterprise software architects to be as vocal as the Slashdot crowd, it would be a really interesting discussion,”
Indeed. However if the TCK was up to snuff, then 5324 variants of J2SE wouldn’t be a problem, coz they’re all compatible. If not, then its a TCK issue.
fwiw, I do understand harmony – but I also agree that forking IS a bad thing on a programming *platform*. It’s bad enough in C++, and thats just a language!
“Why not make something new?”
You’re the one who’s inapt if you can’t think of a single answer because there are a tons of answers.
You are probably one of those hardcore open source fanactics who get upset when someone make a comment against open source related stuff.
You want an answer? I will give you a single one.
Think of Justin Frankel (Winamp) and Shawn Fanning (Napter). We know them because they made something relatively new. We know them because they were creative. But more than everything, we know them because they made something new that we used and loved for years.
Now think about the XMMS developers… Got an answer?
The reason Harmony is silly is not because there should not be an OSS version of Java. It is because there *is* already an OSS version of Java (GCJ + Classpath) that they’ll be competing against and draining resources from.
So you are saying that these projects aren’t innovative:
– Enlightenment
– FreeBSD ports
– Gentoo Portage
– Bit Torrent
don’t generalise…
oh and did I mention Xen?
I never said those aren’t innovative.
Enlightenment: the one and only WM that really got my attention over the years. Why? Instead of trying to rip off MS Windows (like Gnome and KDE), Raster (and Mandrake that left the project few years ago) made something really new. BTW, Raster is an artist. Now we can talk about creativity.
BitTorrent? Another great example of creativity. BitTorrent IS something new.
I didn’t generalize…
It’s just that this:
“Anyway, that’s the only thing most open source people can do. Don’t get me wrong but it’s the sad truth. They are like music bands making cover albums instead of writing their own music. Sure they are good developers but they are not creative at all. Unfortunately, I think that creativity is the most important thing after all.”
kind of gave me that impression…
I said most, but probably I went too far. I’m sorry as well 😉
lets all join hands and sing kumbaya
The only reason I’ve seen put forward as to why there is a need for an open source JVM is for larger platform support; and thus better program portability. Outside of that I don’t actually see a need for an OSS implementation. The convenience of having it pre-bundled in a distribution is not exactly an earth-shattering motivator to write an entire JVM over again when its already freely avaliable for download and usage.
In order for whatever the Apache comes up with to be considered a JVM it has to pass Sun’s compatability tests. Java as a technology will still be very much under the reign of Sun. They will be dictating what the specifications are; and where things will be going. In essence — having an OSS JVM does not give the OSS community any leverage over the direction of Java.
If the Apache team at some point decides they want feature Y in the JVM, and Sun doesn’t like enough to add it to the specification, their JVM, and compatablity testing — guess what – its not Java. Its now a 3rd party extension thats “unsupported”. In essense not really any different than MS “breaking” the portability of Java by making their own modifications.
So I’d have to agree with Gosling. Its not really going to be what people actually want out of it. Its not going to the open source community a say in current or future developement of Java.
Again, alot of open source developers are gonna waste their time by reinventing the wheel.
Anyway, that’s the only thing most open source people can do.
The problem is that Sun’s wheels only work on Fords. Too bad – you have a Chevy. Better luck next time.
It’s not a waste of time for many people still waiting for Sun to make a wheel for THEIR car.
If they had absolutely any interest at all in extending the bytecode, or introducing incompatabilities in the class library, then they wouldn’t be developing a JVM at all. They’d be working on Mono, LLVM, or their own home-grown project and Java as a subject would be largely irrelevant because software written for the platform wouldn’t be able to target the Java platform. They wish to develop an implementation under a liberal license that can be easily retargeted to a number of purposes and support an arbitrary number of platforms in order to liberate developers from dependence on Sun or IBM to provide them with the ability to run Java “everywhere,” and to supply bug-fixes.
If they succeed and their implementation conforms to Sun’s standards, then any derivative work that alters this conformity isn’t going to be able to claim to “be” Java. It does nothing to harm Harmony’s work.
Gosling’s concerns are somewhat disingenuous. There’s nothing that prevents me from developing an incompatible Java-like platform and calling it Bob, or .NET, or what have you. Well I suppose calling it .NET would be difficult, but that’s just because someone else beat me to it.
Who cares if they do not understand, the rest of the world does, in fact SUN doesn not need to understand, just leave people in the OSS camp do their jobs.
“Harmony is complete vaporware. No code had been written.”
False! Fast forward past the diatribe…
“GNU will have to change the license anyway to play with Apache.”
Duh! That’s what it’s all about. Do you honestly think it’s all about redoing Classpath from scratch? If anything, the whole thing is about pandering to “business-friendly” types like yourself who like “free software” as long as “free” is the shortened form of “freeloading”.
“Exactly why is it critical to have a delta between our source licence and the source licence that they think is appropriate?”
I don’t know. Maybe you can ask the authors of the CDDL? :p
I guess you didn’t realize that the Apache license and the GPL are incompatible and that hasn’t been resolved yet. Even with that resolved, Apache has to come up with a VM, flesh out the rest of Classpath. 2 years absolute minimum (more like 3) – and that is highly optimistic with everything going right. Oh, and by that time Sun’s JDKs will be 2 generations ahead, the runtimes will still be better, etc..
f anything, the whole thing is about pandering to “business-friendly” types like yourself who like “free software” as long as “free” is the shortened form of “freeloading”.
No, I’ll be freeloading off of Sun. Why would I even care about some second rate crap that will always be inferior to Sun’s implementation?
Hell, you know the project is already in serious trouble when you have most of the comments on the mailinglist talking about writing the JVM in Java itself. What a disaster.
http://www.jroller.com/page/fate/20050517. You can’t beat that comedy cold when the guy is spot on.
The ASLv2 vs GPLv2 license incompatiblity is being worked on.
cheers,
dalibor topic
Next they’ll be developing Lisp implementations in Lisp, Smalltalk implementations in Smalltalk, and Jikes RVM will exist. Let’s hope the madness can be stopped before PyPy is finished.
Because technical interfaces are not really copyrightable in the real world, see various cases starting with Altai vs. CA.
cheers,
dalibor topic
What purpose is there for this? Overduplication. NOT redundancy. I think we better find out who’s working on Apache’s server. Or better yet, knock off Apache’s design! hahahaha.
Abuse of the OSS liscenses, in an attempt to create a “cheap chinese knockoff”.
There’s software communism, then they’re just outright communism with software. Yech.
Apache, being a wide scale server side implementation, would not really benefit by having their own non-conformant J2SE Implemetation. Though it would have some benefit to be using a conformant J2EE, that Sun and Apache both worked on.
This is mainly why none of this makes sense. Various assertions abound nonetheless.
“Sun suffers from this awful problem. They had a brilliant group of technical people…”
This, IMO, sums up the problem in a nutshell. Sun fired/layed off/downsized/rightsized/etc a large portion of their think-tank a few years back and kept a boat load of PHBs. They have never recovered from this enormous blunder.
Sun’s confused. They have their technology right. Maybe their heart is in the right place too. But they need a proper and clear-cut policy on OSS.
Why?! Why?! I won’t say there is no need for a OS JVM, but seriously, why from Apache!? That makes completely no sense to me.
I’ve been using OS stuff for years. And I’m very happy with a lot of server-side things like Linux, Apache, Tomcat, etc… But frankly, it’s never mattered one bit to me if a JVM came with a particular distro. In fact, even if it did, my projects usually have a baseline required JVM, so I have to download that one anyway.
I develop daily on a Powebook, so I use Apple’s JVM. I deploy on Linux. I have no problems with either of them. I DO, however, have problems with some of the Apache subprojects like Tomcat and Commons. I’m not saying these guys aren’t doing great work, but they have quite a few major bugs! Anyone one who deals with the failed commons logging and Tomcat classloader mem leak issue knows exactly what I’m talking about. And that’s just for starters.
It would seem to me that it would be much more productive to FIX the things broken now. Or even contribute to GNU Classpath. Ya know, first it was Geronimo (which I still think it pointless) and now it’s the OS JVM. What next? The Apache OS?
No one is saying Sun is perfect, but they’ve contributed a lot more code and resources to the OS movement than Microsoft has – I’m sure they have no choice on this one either. Oh, and btw, someone mentioned Novell started Mono. This is incorrect; Mono was started a Ximian which Novell bought. Sure, they’ll shore it up, but they didn’t start it.
Anyway, the whole Apache JVM is a moot argument. Make the other ones out there BETTER. Fix and complete GNU Classpath. Put the efforts to established projects that already have some momentum. Apache is starting to try to be like Microsoft in that they’re trying to be everything to everyone. My advice is to stay focues on their core and make it good. Otherwise, they’ll have buggy or incomplete pieces of everything laying around that no one will want to use.
The link to the article doesn’t work. I want to read it but I can’t!
It sure is a shame that Stallman and others spent so much time developing gcc and gdb and duplicating effort that the folks at AT&T had already put in.
It sure is a shame that Linus and others spent so much time developing Linux and duplicating the effort that the folks at AT&T/USL and all their licensees had already put into creating so many wonderful versions of Unix.
Too bad Brian Fox and others built bash and duplicated all the effort that had already gone into the Bourne and Korn shells.
…and what a shame that the Apache foundation is going to duplicate so much effort that has gone into Java.
What a terrible shame that so much effort has been wasted by companies implementing proprietary products that are useless to the Free Software community. If only UNIX, pcc, ksh, and Sun’s Java had been Open Source to begin with, so much duplication of effort could have been avoided…
Glomek
apache isnt making geronimo because we need an open jvm, if that were the case they would use classpath, which has been worked on for years, has a following, and is a hell of alot closer to being done then anything they start working on now.
No access.
“Who needs Java when we have D anyways?
http://www.digitalmars.com/d/ “
Perhaps because D isn’t yet fully FOSS?
Next they’ll be developing Lisp implementations in Lisp, Smalltalk implementations in Smalltalk, and Jikes RVM will exist. Let’s hope the madness can be stopped before PyPy is finished.
You’re sure out of the loop. There already are LISP implementations in LISP, and Smalltalk implementations in Smalltalk. I’m playing with the 64bit version of Squeak right now (that’s a Smalltalk implementation done in SmallTalk).
You also forgot such things (already out) like Pascal in Pascal (FreePascal), or Ada in Ada. I do believe the gcc C compiler is written in C as well.
So why exactly are compilers written in their own language a disaster? It’s almost universal, not the exception.
The real thing going on with this, the open source Java effort, and the other background effort of an open source implementation of the .NET framework (Mono) is that open source developers are yearning for a free and powerful higher-level language with which to develop next-generation, powerful, and innovative desktop applications.
The idea is that “no one wants to build an aircraft carrier one molecule at a time,” which is what writing GUI applications in C with GTK feels like. In addition, newer developers (like myself) are simply at a loss to contribute anything to the open source desktop because we don’t have the tools that make it easy, fun, and perhaps most importantly, creative.
C won’t go away as the lingua franca of Linux. Linux, as an OS (OS referring here to kernel + GNU), is still largely defined by its rich heritage of command line tools, and most of these tools find C as a perfect-fit language with many performance benefits.
Personally, instead of waiting around for Mono or Java, I’ve been “diving into Python,” which provides the high level you want in a fully open source and cross platform implementation that is available NOW. There is some interesting stuff going on in the Python on Desktop Linux world. For one thing, a developer has created a Python GTK interface designer called Gazpacho, which is essentially a fork/port of a development version of Glade 3. Already, Gazpacho is starting to feel like a “real” UI designer, that allows one to really prototype apps quickly and get work done. It’s only just starting development, and already looks very promising.
Python works well with my existing Linux development tools. I can keep a gvim window open for editing and a console open to ipython, where I can use the interactive shell to inspect the language and test things out live (using features like readline autocomplete and docstring printouts with the ? operator).
Python also has a ton of libraries that are well-tested and clever. I think I’ll be happy with this development environment for the couple years until the Linux community figures out which framework for higher level programming to bank on.
There is a lot of misinformation regarding SUN and their OSS participation. After University of Berkeley, SUN is the second largest contributor to OSS. James Gosling is a brilliant designer (even ESR thinks so) and if you visit his blog, you will find a few small OSS projects of his. Obviously, given that the SUN team has spent so many manhours on *designing* Java, they are unwilling to release it as OSS. But, sooner than later, they would have to release it OSS as they have done (doing) with OpenSolaris. I think developers who have already contributed to Classpath et all will be unwilling to work with Harmony.
There is a lot of misinformation regarding SUN and their OSS participation. After University of Berkeley, SUN is the second largest contributor to OSS.
I keep hearing that stat thrown around, but no one ever seems to be able to substantiate it. From what I know, Sun has opened sourced OpenOffice, put some work into GNOME, released the specs to NFS and (maybe someday) will release OpenSolaris. That doesn’t seem like a lot more than Red Hat who open sourced eCos, Insight, GFS, put work into a list of projects as long as my arm (Linux kernel, GNOME, X.org, FreeDesktop, gtk, gcc, glibc, gdb, rpm, etc) and (maybe someday) will release the Netscape Enterprise Server stuff they bought. So how do you quantify which of those sets of contributions is bigger?
James Gosling is a brilliant designer (even ESR thinks so) and if you visit his blog, you will find a few small OSS projects of his. Obviously, given that the SUN team has spent so many manhours on *designing* Java, they are unwilling to release it as OSS. But, sooner than later, they would have to release it OSS as they have done (doing) with OpenSolaris.
James Gosling, although a brilliant designer, has gone on record as saying several stupid things in defense of Java. I have no idea why, but it’s grating to listen to.
I think developers who have already contributed to Classpath et all will be unwilling to work with Harmony.
If you look at the list of developers who are interested in working on Harmony you’d have seen a bunch of Classpath developers, GCJ developers, Kaffe developers and IKVM developers. Now, that might not stop you from thinking that they’re not interested in helping, but the evidence is stacked against you. Dalibor Topic has posted a few times in this thread, perhaps he can shed more light onto that topic.
Also, regarding the GPL vs APL2, Classpath isn’t licensed under the GPL, it’s licensed under the GPL with several modifications. Talks are under way as to whether those modifications are APL2 compatible, but the ASF is at least trying to collaborate with the efforts already under way.
What’s shocking is why free software hackers are wasting their time on the junk that .NET and Java are. Free software has rich, more productive, versatile alternatives such as Ruby, Python and Perl. It surprises me to see free software hackers fall into the corporate trap that .NET and Java are. Truly sad indeed.
Projects like GCJ are probably more harmful to Java that Harmony ever will be. The fact that you could potentially have developers creating OS/Platform dependent executables from Java sources is a bit more troubling than a fully compatible J2SE implementation. Also, why the hell are they targeting Java 5? Why not go for what’s next, like Java 6. Just an opinion.
Java cannot be freed by an Apache license. Only the GNU/GPL allows freedom. We can not left Apache steal from Classpath. We only need to talk about the GPL
Well just install the bloated crap from SUN …
Gosling is just a silly SUN puppet …
Yeah Swing is better then SWT … pffff
Let IBM finish what SUN started and finally admit that Java has grown to big for SUN …
So why exactly are compilers written in their own language a disaster? It’s almost universal, not the exception
Dude, the guy was being sarcastic.
What I don’t believe is that Java can cut it. Bear is mind that all the languages mentioned are higher level; Java is a special case in that it is probably the lowest level language that targets a VM.
Projects like GCJ are probably more harmful to Java that Harmony ever will be. The fact that you could potentially have developers creating OS/Platform dependent executables from Java sources is a bit more troubling than a fully compatible J2SE implementation.
Why are platform specific builds a problem? If I was going to launch a J2EE app on my servers and I could get a 10% performance boost, why wouldn’t I want to do it? And GCJ can compile to Java bytecode and it includes an interpreter to execute the bytecode, so GCJ isn’t forcing anyone to native compilation.
Also, why the hell are they targeting Java 5? Why not go for what’s next, like Java 6. Just an opinion.
How do you target something that doesn’t exist yet? Sun has complete control over J2SE releases and until they finalize what they want in it you’ll just be chasing a moving target. It’s not like they’re going to be done with Java 5 next week and be wondering what to do, I’d be happy if they can get a Java 5 implementation done by the time Java 6 is released.
Do you suffer from Asperger’s Syndrome or is there some other reason you didn’t understand that I was using sarcasm?
> > Who needs Java when we have D anyways?
> > http://www.digitalmars.com/d/
>
> Who needs D when it sucks so much?
> “åäö”.length => 6
> WTF?!?
Stupid PoS osnews dysfunctional preview-function-wannabe.
“åäö”.length => 6
who needs that when we got BASIC !
Well, on the technical side, one of the ideas being played with on the Harmony lists is to come up with good intra-vm interfaces that would allow us to combine state-of-the art components, but would also provide a way to reuse the work (like new or contributed components) being done within Harmony in other runtimes. Given the goodwill on all sides to resolve the licensing incompatibilities and clarify things between ASF’s and FSF’s license stacks, such work should end up being reusable by any GNU Classpath runtime, and the API designs would be useful in any case among the other GNU Classpath runtimes to energize an exchange of components between them. I, for example, want every GNU Classpath runtime to be able to easily leverage gcj-ed libraries and applications on my system.
cheers,
dalibor topic