Red Hat Joins Open Source Java Project

narramissic writes "Red Hat has signed on to Sun's OpenJDK project and agreed to coordinate its own Java development efforts for Linux with the project. Red Hat will align the work it has done on IcedTea (its own implementation of some parts of the Java SE JDK) with OpenJDK. As part of its participation in OpenJDK, Red Hat will eventually create a compatible OpenJDK implementation for its Enterprise Linux distribution and will also use OpenJDK to create a runtime for its JBoss Enterprise Middleware that is optimized for a Linux environment."

mono

what about mono?

Will anything change for end users?

With all the "openness" going on with Java these days will things get even more complicated? I have 3 important commercial apps that run on java, all three have their own run time environments that are incompatible with each other. I have no end of trouble with jre and firefox. I can't count how many times I've had problems with classpaths trying to run java stuff.

Will the OpenJDK mean another runtime? As in Blackdown, Sun, Open?

Re:Will anything change for end users?

That's not the result of the openness of the JDK or the JVM; the specs for both were always open. Sun always gave you the src.zip for the JDK, and they provided the bytecode spec and in what way to run such bytecode openly and free op charge.

Re:Will anything change for end users?

I think the biggest problem with the original write once, run everywhere promise was this: Java was almost, but not quite, an operating system. If you had a system that worked on WinXP but not Vista, you'd fix the problem; you wouldn't tell your customers to install both operating systems on the same machine because he can't. If I had A,B and C that only ran on Windows versions X,Y and Z respectively, I'd take the trouble to make A and B run on Z.

In Java I can get by with A, B and C each running only on JDK 1.2, 1.4 and 1.6 respectively, but I have to call upon the user, working with the host OS, to deal with any confusion that arises. Now, multiply that by any libraries you use that aren't part of the standard java distribution, and things start to get really interesting.

In part the problem of Java classpaths is one of an embarrassment of riches. There are so many powerful frameworks and amazingly useful little libraries that are out there, free for the taking. However, once you open that Pandora's box, it isn't simply a matter of write once, run everywhere; you have to deal wit the dependency issues that come out of that box. It can be done, and it's not really all that difficult, it's just one of those craft things that some people seem to have figured out how to do and others seem to leave as an afterthought after all the fun stuff has been finished.

The embarrassment of riches also applies to the biggest pitfall in Java development: making bad framework choices, which you can do in multiples at the same time on any project. Leaving aside the case where you have a bad framework (the early EJB stuff -- ugh), or where you have made a bad mistake in requirements analysis, there are so many frameworks that are wonderful for one set of requirements and dreadful overkill for others. Working with a set less than optimally chosen frameworks sucks the joy out of a project, as you set aside the pleasure of solving the client's problem for the tedious drudgery of gluing together mismatched bits of architecture.

parallel universe

I must have slipped into a parallel universe or something because it's starting to look like Java might finally make it's way onto the Linux platform in a useable way. Fair enough we have been deploying Java apps on Linux boxes for a while but it's been much harder than deploying a PHP, C, C++, etc etc application. That always struck me as strange because I would have thought that Java was the perfect language for open source projects. Fairly quick, simple to develop in, stacks of libraries, popular.

Re:parallel universe

The main problem with unresponsive Swing apps, is that most developers do everything within the Event thread, so the app is unable to respond in a reasonable manner.

If Swing developers remember to move intensive operations off the Event thread and into a background thread, then Swing app's are really nice and responsive. It's not that difficult, but for some reason most developers are either unable to, or unwilling to do this simple task.

Believe me, I've seen the source of plenty of Swing app's that have been written with everything in the Event thread and the developer (one of whom I had employed at that time) refused to do this because they couldn't be bothered.

As for the look and feel, it's getting better but it still has a long way to go.

Re:parallel universe

I must have slipped into a parallel universe or something because it's starting to look like Java might finally make it's way onto the Linux platform in a useable way.

Java has been available and has worked well on the Linux platform for years. The main problem in trying to deploy Java on most Linux distros is that the more popular distros started putting GNU Java and GNU Classpath as the default Java VM and classpath on Linux. Sun's Java has been available in binary form for as long I've been using Linux (I started using Linux in 1994). I'm not sure when they started offering source, but it's been available, too, and things like Blackdown have been based on it for a long, long time.

And there are plenty of nice Java apps and environments on Linux -- Eclipse is one of the big ones, obviously. The bottom line is that gcj/gij gave Java on Linux a bad name because standard Java apps and programming examples never have worked on it right. Install Sun's JRE/JDK or Blackdown, and you'll find that Java works great on Linux.

Sun has found religion..

Open Sourced Solaris, SPARC, now Java... Halleluiah cries the OSS choir.

But seriously, this business move by Sun has made it far more attractive to my company, enabling us to test out Solaris on our existing server before we perform a rollout. In addition, having the source code for the UltraSPARC T1 has enabled us to do research into how the chip functions on a lower level, with an eye to further optimizing our software to perform even faster on it. Sun, you might win over my heart just yet.

FYI

The official news on the red hat site:http://www.redhat.com/about/news/prarchive/2007/sun_java.html/ [redhat.com]
because i can't find references on the sun & openjdk site.

I truly hope for the end of gcj/gij

Why? Sure, it was a novell idea to try and create an open sourced java but the whole arguments which backed it up were false. Many people seriously believed that Sun was not opening up the Java source code period, while in fact that was a mere lie. The Java source code was available but simply licensed in such a way which didn't really go well with some. And so they simply declared it "closed source" and fooled many people into thinking that the Java sourcecode wasn't available period.

Why I mention this? Because it was perfectly legal to adopt certain pieces and sniplets of code, check the way things were build an adapting those ideas. All of that might have made a difference for the gcj/gij projects. Personally I condemn those 2 projects, but having said that I will have to admit that they did make a good effort.

But the main reason I hate this stuff with a passion is because its not compatible with Java, and it is my belief that all the nonsense (gcj/gij + the bs about the closed source java) has left Java with a bad name / reputation on the Linux platform. Which I think is unfair and an utter shame. Would this have not been the case I think Java could have lifted some interoperable development movements to higher levels. Sure; it has already done this to some extend and Linux is still a big market for Sun, but when the bs was still spreading you could already easily download binary installers (self extractors) to install Java on Linux. But I have met simply way too many people who had problems to "do java on linux" and when you started disecting the problems it all boiled down to Linux distributions shipping gcj/gij thus resulting in non-working Java software. And as well all know; a good user doesn't blame his tools but the product he's trying.

I once spend 45 minutes on the Sun Java tutorial and couldn't get some examples to work. Eventually I tried on another platform, that did work, and so I knew where to look. Eventually I ended up dumping gcj/gij and replacing it, unfortunately I think many others ended up dumping Java.

Re:I truly hope for the end of gcj/gij

Yes, but a (maybe minor) feature of GCJ is its ability to build self-contained Java applications, i.e. to compile some Java source code into an executable which does not require any previous JVM installed, etc...

I admit that there are few Java applications (at least on Debian) which are compiled by GCJ and packaged as plain old binary executables. Of course, this means avoiding some fancy Java tricks (the dynamic class loader, some reflection abilities, etc...).

Still, I believe GCJ does have at least such a niche market (for those few applications which don't want to depend on a JVM being installed).

Besides, GCJ is GCC based, and GCC is still a nice project (even if it is old).

Re:I truly hope for the end of gcj/gij

Actually, instead of the end, this is just the official beginning: From the intro at [gnu.org]http://gcc.gnu.org/java/ [gnu.org]

Compiled applications are linked with the GCJ runtime, libgcj, which provides the core class libraries, a garbage collector, and a bytecode interpreter. libgcj can dynamically load and interpret class files, resulting in mixed compiled/interpreted applications. It has been merged with GNU Classpath and supports most of the 1.4 libraries plus some 1.5 additions.

From TFA:

Red Hat has signed Sun's OpenJDK contributor agreement and will now align the work its done on its IcedTea project, which was its own implementation of some parts of the Java SE JDK, with OpenSDK, said Shaun Connolly, vice president of product management for JBoss. IcedTea brought together the Fedora project with key Java technologies in a Linux environment, and currently provides open-source alternatives for the few remaining proprietary sections in the OpenJDK project, he said.

Yet looking into the IcedTea project [javalobby.org]:

Red Hat has launched the IcedTea project, with the goal of creating a hybrid fully free Java implementation based on OpenJDK and GNU Classpath. The project replaces binary plugs that are still non-free with code from GNU Classpath "We have been working within Red Hat to replace these binary plugs with free software based on GNU Classpath and to remove the need for bootstrapping with unfree software. This is important for a number of reasons, the most pressing being that only free software may be used to build operating systems like Fedora", said Andrew Haily on an OpenJDK newsgroup.

Also, Wikipedia [wikipedia.org] references "Wielaard, Mark [klomp.org] (2007-06-07). IcedTea [wildebeest.org]. Retrieved on 2007-06-09":

IcedTea replaces the binary plugins with the equivalent GNU Classpath code, compiles it all using GCJ and optionally bootstraps itself using the HotSpot Java Virtual Machine and the javac Java compiler it just built.

So again, this is not the end of end of GCJ but part of its validation.

Other Linux Java Options?

My problem with the Sun JRE is that it is HUGE. Why do I need 100MB+ to run a simple Java application?

Are there other good JRE options for Linux? Maybe something geared towards embedded environments?

Re:Other Linux Java Options?

That number is a bit exaggerated, my install of the latest Java 6 JRE is about 80MB (and the download is only 14MB).

One of the reasons it's so big is because it has a LOT of functionality. But you're right of course when you say that you don't need all of that to run a simple Java application. So Sun decided to do something about that: in the upcoming Java 6 Update N (what was previously called the "Consumer JRE") only a relatively small "kernel" will be installed which has only the most essential components. The rest will be downloaded "when needed".

Re:Other Linux Java Options?

My problem with the Sun JRE is that it is HUGE. Why do I need 100MB+ to run a simple Java application?

you don't. a simple stroll over to java.sun.com will show you that the JRE [sun.com] is 14M for windows and 18M for linux.

the "100M+" is if you're also downloading all their development tools and documentation (and possibly netbeans, depending on the link). not atypical in the least.

Re:Other Linux Java Options?

What does the uncompressed local copy have to do with download times? 14MB compressed takes just as long as 14MB uncompressed. If you think that your CPU can't handle fast decompression, just think of all of the web sites that gzip their content for network efficiency.

As for the complaint about docs, are you serious? Are you seriously complaining that there is too much documentation available in HTML format? And optional documentation at that? Think about what you're saying for a second: that you consider it a drawback that every class, method, and member of the JRE is consistently documented in detail.

GUI: AWT versus Swing are native widget peers versus internally rendered widgets.

RPC: RMI, CORBA, and XML-RPC/SOAP are for the following in order: RPC in a 100% Java environment, cross-platform binary RPC, and XML text-based RPC. There is a place for each of those.

XML parsers: are you referring to the SAX, DOM, and StAX parser APIs -- which would make three? Or do you mean two parsers like Crimson and Xerces. I think the former is self-evidently a good thing. The latter is due to compatibility and consistency through multiple releases as the older parser behavior may be necessary for an older app even if it's a little slower or more memory inefficient.

I can see your argument against including a scripting language, but Sun wanted to include a reference implementation of their pluggable scripting interface.

I/O: Blocking vs. non-blocking. What's the problem? Both have their uses.

What you call bloat, some would call completeness. Let's compare against some other popular languages.

Common Lisp: 10MB
Latest Python download for OS X: 17.9MB
Latest Perl download for OS X: 33.5MB (Linux version is between 18.9 and 24.8MB)
Latest Ruby (without Rails) download for OS X: 13.71MB

But don't take my word for it. Download for yourself. The only reason these other languages seem smaller to you is because they are bundled seamlessly with your Linux distribution.

Want database access, RPC, non-blocking I/O, XML parsing, etc. from those languages? Too bad, that's another download. Sure there are resources like CPAN, but why are their cores so bloated? Somehow Java is able to provide all of those "bloated" APIs at about the same download size as those languages that lack them.

And don't get me started on C and C++. They don't even have a standard database layer, XML library, or the like for you to download separately. Learned one non-blocking I/O library? Too bad, your new company uses a different one. Do you think ODBC is a good solution? Obviously you've never programmed for it.

I'm sure I could go on, but you get the picture.

Getting harder to choose: Java or Python?

How do you folks intelligently and objectively choose between Python and Java? Or at this point is it a subjective decision?

Re:Getting harder to choose: Java or Python?

They each have their advantages and disadvantages. According to the great computer language shootout [dada.perl.it], Java is faster by an order of magnitude, but is more verbose and usually consumes about twice as much memory. The ultimate decision depends (at least) on your application's needs and the capabilities of the environments you're targeting.

this will finally means java for PPC :-)

RH interest in OpenJDK has some interesting side effects: they're actively working on porting it also to powerpc, both 32 and 64 bit.
Gary Benson, main RH powerpc developer, is keeping a journal [livejournal.com] about his work, quite interesting stuff.