Will Sun's Java Go Open Source? 519
Ritalin16 writes "CNet report that Sun Microsystems wants to send Java closer to the open-source world, yet keep it safe from harm. "Project Peabody" adds two licenses that make it easier for outsiders to see the code. But Sun stops short of embracing open-source. Sun's licensing practices for Java are closely watched. Proponents of making Java open-source argue that a different license and development process will help accelerate usage of Java, which faces ongoing competition from Web open-source scripting tools, such as PHP, and Microsoft's .Net line of tools."
Re:Sun's tight control... (Score:4, Informative)
Re:What's so significant? (Score:3, Informative)
An irrelevant question, since Sun has no plans to open-source their VM.
Since the Java specs are anyway open, is there any point if Sun opens up the JVM implementation?
Writing your own implementation is really, really hard. Especially since the "spec" doesn't specify everything. So in the near future Sun-derived VMs will be the only complete ones.
Re:Integrate into FireFox? (Score:5, Informative)
It is exploiting a clueless user that doesn't know enough to click "No" when presented with a warning dialog that says that an untrusted application is trying to make changes to the local filesystem, and it asks permission from the user if this should be allowed. Previously, it had been assumed that a user would be ever so slightly clueful enough to not actually permit an untrusted application to do this, especially since they are getting a nice and friendly warning dialog, but hey... I guess user stupidity reaches new lows every week. The people that are vulnerable to this are the same people that think it's a good idea to run native exe attachments in emails from unverified sources.
Get informed (Score:5, Informative)
Re:off-topic-a-roony (Score:3, Informative)
I don't know of anything beyond that.
The community would benifit from having a Open Source java implimentation because it would allow distributions to distribute java runtime enviroments and allow tighter integration of java into general linux offerings. Like you could begin writing Java-based GUI applications and then distibute them yourself without requiring your audiance to agree to restrictive licenses and download and install Java on their own.
Probably a bit to late for all that, unfortunately. Sun had it's chance and now Linux has managed code in the form of Mono (open source
Then you have stuff like Python, which is a very mature, powerfull, and easy to use general purpose language, and PHP which is very commonly used for web scripting.
If they did release a OSS version a few years ago java on Linux would probably be the defacto way to write GUI and Web applications.
Not saying that it would ROCK or anything, but the restrictive licensing realy held Java back.
Re:Sun's tight control... (Score:2, Informative)
Why do you think that? IBM created a superior background garbage collection for example and gave it back to the community. IBM of course write JVM's for their own OS's, and those JVM's meet Java Community specs.
It would do IBM no good to have a proprietary JVM. I haven't even heard of proprietary Java extensions from anyone but Microsoft. Would you know anything about that?
rd
Re:zerg (Score:3, Informative)
If you want to be able to do something with the code and distribute the end result you should look at "Free Software".
By definition Free Software is Open Source but not necessarily vice versa.
Re:Already ditched (Score:5, Informative)
next?
(yes, I know it's still more typing than python, and I know you were tying to be funny, but Java really isn't that much work)
Re:Why do colleges proselytize with java? (Score:2, Informative)
We've already been down the Python/PHP "typeless" road with teaching Visual Basic, and the end result is a lot of sloppy programmers.
Re:Sun's tight control... (Score:2, Informative)
And IBM's AS/400/iSeries/i5 is 64 bit PowerPC, so the Java JVM on the AS/400 would be 64 bit.
rd
Re:Speaking of Java..anyone know (Score:5, Informative)
Anyway, some of the most popular would probably be JBuilder [borland.com], JDeveloper [oracle.com], IntelliJ IDEA [jetbrains.com], KDevelop [kdevelop.org]...
If you prefer more light-weight IDEs, you can always use ANT together with something like Emacs or JEdit [jedit.org].
Re:off-topic-a-roony (Score:4, Informative)
Regards,
Steve
Re:Already ditched (Score:3, Informative)
public static void main(String[] args){
}
public void print(Object o){
System.out.println(o);
}
Makes everything I write just that much easier
Kleedrac
Re:No, they want to keep their integrity. (Score:4, Informative)
Re:Integrate into FireFox? (Score:1, Informative)
Re:No, they want to keep their integrity. (Score:5, Informative)
By definition of the term `open source' as it is used now, it means that people can get, change and use the source and at least with some restrictions redistribute the resulting changed version.
I.e. `open source' more or less means `you can fork'.
The solution is perhaps to grow up and stop wanting to keep control. Life is to short to worry about that kind of thing.
And, of course, the fact that control is not enforced by the licence doesn't mean that there will be no central control. I think most people would say that Linus has reasonable control of the future of the linux kernel, dispite the fact that anyone could fork it, and many do for special purposes.
Consider XFree86 vs X.org for an instance of why the ability to fork is a good thing. Sometime when you get bored with working on your library, or fall under a bus, someone else with enthusiasm can take over, and if they do a good job, their branch will take over from yours and, assuming it wasn't a bus, you can go off and do whatever you are more interested in doing at that point.
Re:and how's that working out? (Score:5, Informative)
Ok. I'm thinking of perl. I'm thinking about various systems I've seen and worked on over the last couple of decades. I'm thinking "fairly popular" is a pretty serious understatement when talking about perl. Tcl/Tk? Maybe not so much, but I've still seen lots and lots of it running behind the scenes on all sorts of systems. Python? Ok, I haven't seen much enterprise penetration there. Yet. But it's pretty new, relatively speaking. And it's already got zope, plone and gnue, and looks like it's only going to be growing from there. (Anyway, I can hope, 'cause python sure seems to me to suck less than perl, java, or tcl.)
Microsoft has cloned Java
From what I've seen, they haven't merely cloned it, they've actually improved on it. I know, a blasphemous thing to say on
I think you have to explain your reasoning a bit here.
Ok. I didn't think it was that tricky of a thought, but since you ask.... Kaffe and gcj are not forked from each other, they're forked (very loosely speaking) from the design of their not-open-source predecessor. Nobody is trying to fork kaffe or gcj - they're already open source, so there's no particular reason to fork them. And if java had been open source from the start, there never would have been any reason to create these competing projects.
Ask anyone involved with either project (kaffe or gcj) if they would have even considered working on those projects if java had started as open source! If java had been open source there would be no kaffe or gjc! Period! I don't know how I can explain it any more clearly than that!
And, for that matter, if java were open-sourced, even at this late date, I would expect to see kaffe vanish almost instantly, and I wouldn't place any bets on gcj's long-term prospects.
GJC (Score:5, Informative)
However, the biggest problem with the GNU compiler is the lack of compliant class libraries. The GNU Classpath project, which aims at creating an open source cleanroom implementation of the Sun class libraries are severely behind. My optimistic guesstimate would place them at being about 3 years from full 1.5 compatibility, at which time we're probably using the beta of 1.7.
However, GNU Classpath has an easier job ahead of them than the Mono folks, which begs the question as to why they chose to go with the Microsoft technology in the first place.
Re:Already ditched (Score:2, Informative)
// Import of static methods and fields
import static java.lang.System.out;
...
private void someFunction(String someVar) {
out.printf("Hello %s\n", someVar);
}
Re:No, they want to keep their integrity. (Score:3, Informative)
While I do not disagree with your argument on why Java needs to be protected, your statement that there is no formal spec for
From the microsoft site:
The Common Language Runtime (CLR) is Microsoft's commercial implementation of the Common Language Infrastructure (CLI) specification. The CLI specification is an international standard (read ECMA) for creating development and execution environments in which languages & libraries work together seamlessly.
Read More [microsoft.com]
The C# language is also an ECMA formal spec.
For a serious alternative implementation, Go-Mono [mono-project.com]. As to its unclear legal status, we have our own Miguel De Icaza discussing these issues, Here. [ximian.com]
I guess he is tired people are raising the same issues everytime. Its been done to death. And Miguel is quite liked inside Microsoft too.
Re:gcj is nice (Score:5, Informative)
Actually, yes. Red Hat has actually aliased 'java' to them in the default install for quite some time now.
The more interesting thing is that FC4 will for the first time include natively-compiled Java packages. And not just any packages, but Eclipse and Apache Tomcat.
Re:Already ditched (Score:2, Informative)
I see you never did any plain text parsing in Java. Sure, Java has regex. But you'd be amazed how compact plain text parsing code is in Perl.
OTOH, I like Java over Perl for it's stringer typing and for it's usable OO environment (usable 'cause multiple inheritance is missing and 'cause I dislike misusing interfaces for this.)
Re:Java is open like C (Score:5, Informative)
Wrong. There are many implementations of C. There are three of C# (excluding the runtime).
There is no need to reverse engineer anything since JavaDoc spells it all out for gou,
No it doesn't. Sun's documentation is terrible in lots of places.
and heck Sun provide the source code to the files too!
Looking at Sun's sources if you want to make your own implementation is legal suicide.
It bugs me that people perfer Mono'a C# over Java because it's "more free".
It IS more free (see below).
If they spent half the time coding a JVM that they've spent coding a Mono they'd be done years ago.
Well, I would certainly like to to see more resources go towards free Java development, but the missing thing at the moment isn't a VM. There are lots of free JVMs (gcj,kaffe,jamvm,sablevm,jikesRVM,cacao and so on), some of which are quite mature.
The missing part is the class library.
You can use IBM's or Apple's or your own.
These are not seperate implementations; They all use Sun's code, notably the class library. They are all covered by the same licensing restrictions set by Sun.
Is there some big piece that I'm missing that would bother anyone besides GPL Zealots?
Yes. For instance, what if you'd like to include the install of the runtime in your own installer? Can't do it.
What if you do embedded work and would like to exclude the parts of the runtime you don't need in order to save space? Can't do it.
What if you do find a bug in Sun's code, and include a work-around in your code? Bravo. You just violated Sun's license.
Re:Already ditched (Score:2, Informative)
Class.method(argument) do |arg1, arg2, arg3, etc|
Also available is {} syntax, like:
Thread.new(session) { |my_session|
One stupendously fantastic thing about blocks like this is that the callee can then do any necessary cleanup. The caller doesn't have to know how to terminate a thread. It just falls off the block, and when it does so, the callee does whatever is necessary. Used very frequently with files, as:
File.open(filename) do |file|
foobar = file.read
end
So inside the block, "file" is a new File object that can be treated as you would any other File object. Falling off the block automatically closes the file for you, so you don't have to worry about it.
It took me a couple tries to realize how wonderful Ruby's syntax is, but it didn't take long for the "holy shit" factor to kick in, and I was off and running. You don't really appreciate being "fully OO" until you actually try a language that is.
For me, Ruby is just about the perfect language; certainly better, from a purely syntax-oriented angle, than anything else. Ruby needs more modules/libraries/gems and could be faster (hopefully the Ruby 2 VM will help with that), but it's still outstanding as it currently stands.
Re:off-topic-a-roony (Score:3, Informative)
Having seen some of the simplistic negative responses to this post, I have to say it's not that simple. gcj has its strengths and weaknesses, and for some things it's substantially faster than the commercial Java VMs.
Performance wise, its weak spot seems to be GC performance and some other specific areas like synchronization monitors and exceptions. However, if you tailor your code to gcc's strong points, by using techniques like object pools, you can achieve very strong performance. If you can avoid on-the-fly classloading (and thus any interpreted code) you can get excellent performance and the added benefit of totally deterministic execution speeds. This is good for real time code - dynamic runtime optimizers are scary for real time applications, as is GC in general (another reason object pooling is a nice technique).
Hope it was interesting...
Re:off-topic-a-roony (Score:3, Informative)
From www.blackdown.org:
Re:Java is open like C (Score:2, Informative)
That's not quite true though, because there is: GNU Classpath. Which is used by just about every free VM out there. Kaffe and GCJ are major contributors to Classpath.
Repeating myself, the class library is the major issue. It's much, MUCH less work to write a VM than to implement the entire class library. Also, the VM changes less than the class library; new packages get added a lot more often than new language features.
Not good enough (Score:1, Informative)
And royalty-free RAND is not necessarily open source compatible. More than one Microsoft product has been licensed as "you use this library for any purpose except on open source applications" or "you use this library for any purpose but only on a Microsoft licensed operating system".
The biggest problem for sun is... (Score:3, Informative)
Look at what happened with microsoft, they took the sun JVM and did an "embrace, extend and break 3rd party compatibility" act on it and so Microsoft (up until sun forced Microsoft to remove it) was shipping a JVM that was incompatible with the official VM from sun (and people actually built java apps that only work on the Microsoft VM)
What sun should do (IMO) is to release the code under an Open Source licence. BUT, if you want to use the sun java trademark (e.g. to represent that what you have released is a java VM or whatever), you need to pass a compatibility test and get a stamp from sun. Then, they could release the compatibility test to the world so that everyone can test their code before it goes to sun for approval. The same test could apply to any java VM or compiler (e.g. GCJ).
People who want to modify (and share modifications to) the JDK, JRE and JVM can do so. (e.g. people wanting to port the JRE/JDK/JVM to a new platform/OS)
People who want to try out "unapproved" modifications can do so (with the full knowledge that they arent official and may not be 100% compatible with the official sun releases)
People who want "java" so they can run java apps can feel safe and only download stuff approved by sun.
If the licence was like GPL where source must be
released if binaries are released, it would help prevent what happened with Microsoft and the MSJVM (since any incompatible modifications would mean that you wouldnt get approval and couldnt call it "java" and also any modifications you make would need to be returned to the community).
Because of the rules for the use of the java name, there would be every incentive for anyone making changes to the JRE/JDK/JVM to do the compatibility test and get their changes approved (the licence, the trademark licence and the other factors would work against any incompatible forks showing up especially since companies like MS with an interest in incompatible forks would not want to use code under a licence like the one I suggest)
Developers would have a big incentive to write 100% compatible java code (since they would be instantly able to see that their development setup is not "compatible" and unless there is a good reason to use that incompatible VM (e.g. like what happened with the MS VM being used because it was shipped with windows, IE etc), they would be more likely to choose one that had the "java" name and sun "seal of approval".
And by having a dual licence and/or a "copyright assignment" like OpenOffice/StarOffice, they could also continue offering commercial licences for java like they do now (just like they use outside code contributions to OpenOffice in the commercial StarOffice product)
There are probobly holes in my plan somewhere through, something I havent thought of.
Re:off-topic-a-roony (Score:3, Informative)
Now, it may be that the JVM simply takes 2 seconds to start up and the GCJ version doesn't. I do not know. However, I have no complaints about the performance of GCJ. The biggest problem with it is simply compatibility: quite a lot of Java APIs don't run on it.
Re:off-topic-a-roony (Score:3, Informative)
Re:Why is there a problem seeing the code? (Score:3, Informative)
At my previous employer (a big IT company you'll have heard of), we were not allowed to even LOOK at GPL software without going through an approvals process first: for fear of "infecting" proprietary code with GPL methods.
I've not looked at these new Sun licenses, but maybe the idea is that a corporate legal team can look at them and far more easily say "sure, we're happy for you to look at that source".
You can't study Sun's code to learn from it (Score:5, Informative)
Gentoo Java fact I was very impressed with (Score:3, Informative)
Besides supporting bunch of JVMs, Gentoo integrates all the Java stuff in Gentoo way: you can just "emerge jboss" and you will end up with a complete server environment with all the scipts, etc. Very nice.
Re:off-topic-a-roony (Score:5, Informative)
C# !=
Not even close.
Re:off-topic-a-roony (Score:3, Informative)
While I'm not sure how SUSE manages to distribute Sun's JRE, Gentoo doesn't distribute it at all. Instead, the ebuild merely simplifies the process of downloading the file from Sun and installing it.
While this may seem like a small nit to pick, the fact is, if you look on the Gentoo CDs, they distribute the Blackdown JRE, not the Sun JRE.
Re:off-topic-a-roony (Score:2, Informative)
Re:Already ditched (Score:2, Informative)
I don't care about local method variables having types declared, but to me it makes a lot of sense for method arguments and return types.
And before you flame me, you better go read Guido's blogs over at artima.com because that's one of the reasons he's going to put in optional type declarations in Python 3000.
Re:Already ditched (Score:3, Informative)
Re:Sun's marketing clowns (Score:2, Informative)