Learn about Linux
Download Linux
Get Linux helpSpecial Offers
Feature: Java
IBM, Sun meeting to discuss open source Java
Share
Print
CommentsSun said Thursday it would need a few days to consider IBM's challenge proposal and provide a complete response. However, things have been changed, and a meeting has now been arranged between key decision-makers of both companies.
Even though the high-level discussion may have begun, Sun will release no official statement on it today, a spokeswoman told NewsForge.
In an email message to Gingell Wednesday night, IBM evangelist Rod Smith challenged Sun to help create an independent group to develop an open source implementation of Java -- something both the Java and open source development communities have been requesting for years.
"Rob takes what Rod says very seriously, so he's changing his schedule in order to get a response out as quickly as possible, but we're not going to have an answer for such a key strategic issue as this in 24 hours," Sun's spokeswoman said.
Smith -- who used to work for Sun -- cited a recent eWeek article in which Sun evangelist Simon Phipps -- who used to work for IBM -- asked: "Why hasn't IBM given its implementation of Java to the open-source community?"
"Simon's comment appears to be an offer to jointly work toward this common goal," Smith wrote in the email message. "IBM is a strong supporter of the open-source community, and we believe that a first-class open source Java implementation would further enhance Java's position in the industry by spurring growth of new applications and encouraging new innovation in the Java platform."
IBM and other companies -- not to mention many enterprise developers who work in the open source community -- have been asking Sun for years to open up Java, mainly for the reason cited above: encouraging innovation. Other journalists and analysts believe that an open-sourced Java will give Sun much more mileage than that; by releasing its No. 1 technology, they say, Sun could completely reverse a negative image to developers -- mainly for its difficult, and, some say, one-sided licensing practices. This would improve its public relations and put it far ahead of chief rival Microsoft in the eyes of developers.
Smith also wrote: "Sun's strong commitment to open-source Java would speed the development of a first-class and compatible open-source Java implementation to the benefit of our customers and the industry. IBM is ready to provide technical resources and code for the open-source Java implementation while Sun provides the open-source community with Sun materials, including Java specifications, tests and code. We are firmly convinced the open-source community would rally around this effort and make substantial contributions as well."
Creating an open source Java would almost certainly speed up the adoption of Java-based Web services and service-oriented architecture. Java already is the de facto enterprise platform; the Gartner Group and IDG both estimate it to be used in 60 percent of all enterprises. An open source boost would really send it up the charts.
"I am convinced that the creation of an open-source implementation of the Java environment would be of enormous importance to the developer community and our industry's collective customers," Smith wrote. "It would open a whole world of opportunity for new applications and growth of the Java community."
Sun's first answer to this challenge three years ago was the Bill Joy-championed Project JXTA, a set of Java-based, open source peer-to-peer protocols that allows any connected device on the network to communicate and collaborate with each other (such as PC to PDA, cell phone to PC, PDA to server, etc.). The JXTA community of developers numbers about 16,000, Sun said.
JXTA is a subset of Java, but it is not Java. At the time, the move was applauded by the open source community, but there was some residual frustration at Sun.
If history is any indication, Gingell will reiterate the standard company line: Java is the Sun software franchise, and you don't give away the franchise. However, if Gingell and Sun decide to do something radical and agree to open source Java, it would mark an important turning point in the history of the 22-year-old company.
Smith's email was just the latest in a series of "open letter" email challenges that key executives at the two companies have been sending.
On January 21, Jonathan Schwartz, Sun executive vice president for software, posted an open letter to IBM on Sun's Web site, inviting Big Blue to purchase Sun's Linux-based Sun Java Desktop System for IBM employees.
Schwartz said IBM, in an internal memo, challenged its own IT organization and all of the company to move to a Linux-based desktop before the end of 2005, so he offered the existing JDS as a solution.
Comments
on IBM, Sun meeting to discuss open source JavaNote: Comments are owned by the poster. We are not responsible for their content.
I hope that we'lll get the CLI that<nobr> <wbr></nobr>.Net has - So I can compile my C/C++/python/php code into a module and link it in without platform/compiler issues
Maybe if they open source java they will include some of the tool kits, and then the gcc java compiler runtime will be able to pick up some of that code and give you your desire
> I can compile my C/C++/python/php code into a
> module and link it in without platform/compiler
> issues.
This *is* exactly Perl 6. People miss the point that Perl 6 and Parrot (the VM) will change the whole landscape. Parrot is meant to be the VM to run Perl and any other language, including hopefully Python and Ruby. Once Perl 6 and Parrot debut, it will only be a matter of time before a translator for Java bytecode to Parrot bytecode comes out.
Imagine combining the expressiveness of Perl with the structure of Java. Once these disparate languages work together on Parrot, we can return to the age-old Unix adage, "the best tool for the job". Then we can get more work done, rather than endlessly reproducing efforts in various languages when there is no benefit.
Is this what Parrot is all about?
Pyhon Bytecode --> Perl --> Execution at runtime
Java Bytecode --> Perl --> Execution at runtime
Ruby Bytecode --> Perl --> Execution at runtime
In other words, you make an interpreter for which you make another interpreter for which you code your program for.
I thought Java was slow on its own, whitout another interpreter interpreting its own interpreter.
Wait! I got an idea. Lets write a Perl VM in Perl! That way Perl would really be ported to every platform. Who cares if that second VM would only run if you have a machine-os native VM and it would be really slow to run real world programs. Better yet, lets run our VM in another copy of our VM in another copy of our VM in another<nobr> <wbr></nobr>... in the real VM in some actual machine. Yes! that's it.
Or better yet, let's rewrite <A HREF="http://bochs.sourceforge.net/" TITLE="sourceforge.net">Bochs</a sourceforge.net> (a program wich emulates an i386/i486/Pentium/Pentium Pro/AMD64 based machine at the hardware and bios level so it can run multiple OS and it is written in C++ so it can be compiled for other hardware platforms) in Perl! That whay we can have all on that running on top of Bochs running on an aditional copy of our VM running on<nobr> <wbr></nobr>.. another copy of Bochs running on another Bochs running on an a really old Amiga!
Talk about SLOW.
<SERIOUS>
Actually, a Java to C++ compiler would be a more interesting project, if Sun decides to open source it.
</SERIOUS>
Perl source --> Parrot bytecode --> Parrot VM
Python source --> Parrot bytecode --> Parrot VM
Java source --> Parrot bytecode --> Parrot VM
BASIC source --> Parrot bytecode --> Parrot VM
And so on. Compilers for all these languages would be re-implemented to generate Parrot bytecode rather than bytecode for their respective native VMs. It's basically the same strategy as the<nobr> <wbr></nobr>.NET runtime, except with an open-source, register-based VM instead of a closed-source, stack-based VM.
It can certainly be done; Python has already been re-targeted for the Java VM ("Jython"); and Parrot is already running BASIC and partial implementations of Perl, Python, and several other languages.
You claim to be a seasoned C/C++/javascript developer and were not able to learn Java?
You hate the _language_ because the _UI_ is dated? What percentage of software written in Java does require a UI, what is your guess?
You dream of using the NT Kernel with KDE instead of the Linux kernel?
A<nobr> <wbr></nobr>.net'ized Java will allow KDE on one runtime?
Beg your pardon?
>> Linux's charm is it is the same on all platforms (well close enough to not matter much at all)
What do you mean 'on all platforms'? If you mean h/w platforms, I agree to some extend.
NO its not. Linux is a kernel. I think he meant run GNU on the NT kernel? Or GNU/Gnome? No wait, the G in Gnome is GNU. Ho hum. So confused.
I'd like a "Desktop Linux" distribution to run on my "NT kernel" until the hardware people "get with it" and "start distributing and supporintg their hardware drivers along side windows."
Think about it: NT GNU Desktop. The best of both worlds.
>box. I can do it now via Cygwin, but it is
>dreadfully slow.
Well, it still will be slow, because you're going to replace one emulator with another. (Cygwin with Java)
Possibly this problem could be resolved with faster PCs... Pentium 5, next Athlon,<nobr> <wbr></nobr>....
By the way - I'm running KDE under WindowsXP via Connectix VirtualPC 5.1 - and it's pretty fast.
Ummm... Did you have a labatomy and have to start coding JavaScript? Get real. Any 'seasoned' C/C++ developer finds Java easy to learn. I have to guess you are a JavaScript coder who has done hello world in C/C++. If you can't fingure out AWT/Swing you certainly can't figure out MFC.
Its destinity not being tied to a single vendor will allow it to fully become a reliable standard.
More compatibility between Lotus and OpenOffice.org!
It would be superb if Lotus adopted the OOo file formats as its defaults, hugely increasing global interoperability of office suite software and advancing free file formats significantly. I see a lot of potential here...
Yay!
Posted by: Scorp1us on February 27, 2004 11:19 AMBut things completely changed. I hope that we'lll get the CLI that<nobr> <wbr></nobr>.Net has - So I can compile my C/C++/python/php code into a module and link it in without platform/compiler issues.
This alone would set a revolution in motion. We'd have standard "assemblies" (not to borrow from MS, but it is a good term) that will be the foundation for computing. We'll have a completely cross-platform platform. My life as a developer will be good.
This unfortunately will steal some wind from the sails of Linux. Linux's charm is it is the same on all platforms (well close enough to not matter much at all). Window's NT kernel is stable 'enough'. We'll kill off the drive for linux on the desktop.
One of my dreams is running KDE on my NT kernel box. I can do it now via Cygwin, but it is dreadfully slow. Having<nobr> <wbr></nobr>.net'ized java will allow KDE on one runtime. OSS apps will blossom, and so will stability (cause most crashes are app crashes). Linux will stall on the desktop with this ne found stability. Linux on the server will continue though because cpu cyles count more there.
I forsee my dream coming true. I'll have 'linux' on windows with completely working WINE - because it is no longer emulated, it just is. I won't have to worry about drivers, since I'll be using the winnt drivers.
Indeed EVERYONE even MS stands to benefit. But MS will have a price.<nobr> <wbr></nobr>.Net will die. Without having the openness, they won't have the adoption. As I read today, MS is hemmoraging. It's only going to get worse. Dump your MS stock now, buy IBM and Sun.
#