Will Sun Open Source Java?

capt turnpike writes "According to eWEEK.com, there's an internal debate going on at Sun whether to open-source Java. (Insert typical response: "It's about time!") Company spokespersons have no official comment, as might be expected, but perhaps we could hear confirmation or denial as early as May 16, at the JavaOne conference. One commentator said, "Sun should endorse PHP and go one step forward and make sure the 'P' languages run great on the JVM [Java virtual machine] by open-sourcing Java." Would this move Java up the desirability scale in your eyes? Could this be a way to help improve what's lacking in Java?"

Comments

If they do, it will all depend upon the license.

[khasim]

"Open Source" covers a LOT of licenses.

What changes and how would depend upon which license was chosen.

No

[aliquis]

"Will Sun Open Source Java?"
No, haven't they already said that? Like hundreds of times? And does it really matter?

"Sun should endorse PHP and go one step forward and make sure the 'P' languages run great on the JVM [Java virtual machine] by open-sourcing Java."
"No", who would run PHP on Java anyway? Why? Why would open-sourcing it help?

"Would this move Java up the desirability scale in your eyes?"
No, Java is already desirable in my eyes.

"Could this be a way to help improve what's lacking in Java?"
No, what is lacking?

People who complain that Java is slow, should be open-sourced, and so on have never seemed to had a clue.

Re:No

[Enonu]

Hear Hear!

Will people stop trying to move Java towards a culture that won't keep Java up to the same standards Sun has? There's a reason why the top two server side platform these days are .NET and Java, and it's because a there's a quality standard and completeness not found elsewhere. The next time I look at another configured by altering it's code, hard tied to MySQL, non-tiered POS LAMP application, I'm going to cry.

The only place I ever see Java going is perhaps to be bought by another bigger company who has a similar path. My only hope is that it's IBM because their Java apps are of a higher quality than Sun's, and they've done such good work with the Eclipse platform.

Re:No

[JanneM]

People who complain that Java is slow, should be open-sourced, and so on have never seemed to had a clue.

Irrespective of any ideological issues, there are a few reasons the current situation hurts Java a bit.

Foremost for quite a few readers of slashdot is that free Linux distributions can't include Java in their default install. That means Java-based apps are not going to be included either. And since users need to jump through quite a few hoops to get Java installed (don't say "it's easy" - for most people anything beyond using their package manager is too high a hurdle), you can't assume it will be available on desktops in general.

The second issue is that Java does not really play well with the desktop. I have set up my desktop to run fine using three languages - English, Swedish and Japanese - and made sure everything from localization to character input works smoothly. But Java does not cooperate; it has its own way of dealing with CJK characters and needs its own fonts and separate setup to work. I have fiddled a little with it, but have never gotten it to work properly (especially being able to run an app in Swedish while still being able to input Japanese). And since it uses its own input method, it does not share the local dictionary so typing becomes frustratingly different from any other application I use. And since the code is not open, distributions can't fix these interoperability issues.

Both of these issues serve as disincentives from using Java apps and from writing them in the first place.

Re:No

[TheWama]

How about the people who claim it's tedious and sprawling*?

Simple example...

Java: BufferedReader in = new BufferedReader(new FileReader("foo.in"));
Python: in = open('foo.in')
Ruby: in = File.open('foo.in')
C++: ifstream in; in.open("foo.in");

To be fair, I'm no Java expert, but in my experience with it, I'd have to be masochistic to look at it all day...

Re:No

[Cederic]

Sorry, but when I'm dealing with code written offshore by developers of mixed ability in a system integrating my mainframe to my fulfilment systems with customer web access, online payments, content management and back office consolidation thrown in then I want a language that's immediately readable.

Perl is exceedingly brief to write. It's also very difficult to read.
PHP/Python are better, but most code I see is frankly horrible.

With Java the APIs and libraries - even third party, even open source - have descriptive names. They make sense. You can figure out what's going on from the method and classnames, without having to go and read them.

This makes unfamiliar code easier to read. I've done Delphi/C/C++/perl/PHP programming, and none of them have as easy to read code as Java. I value that a lot.

That you have to type an extra 32 characters to open a file? Good! Hell, you're not typing enough already - 'in' is a bad name for a variable. How about making it descriptive, tell the reader what you're reading.

Modern IDEs do most of the typing for you. The clarity of the code that's written is more important than the extra two seconds you need to write it.

Re:No

[shish]

Why do none of your other language examples have an explicit buffer layer? You should add that, then see what the result is :P People in other languages have become used to reading chunks of a file into a buffer then parsing a bit, then reading a new chunk -- java people can just read what they want, when they want; IMHO adding a bufferedreader is a small price to pay to get rid of all the manual buffering~

While you're at it, make it read objects from a gzip compressed network stream:

new ObjectReader(new GZipReader(new SocketReader(new Socket("1.2.3.4", 42))));

(I'd really like someone to demonstrate this in other languages; having started with java and moved to C, things like gzip stream reading and network access seem like such unnecessary pains in the ass that I've never bothered to learn them properly...)

Alternate VMs

[Flounder]

Could an open source be faster, more stable, and have better resource management than the Sun VM?

Certainly couldn't do worse.

Save Apache some time ...

[SickLittleMonkey]

... and just do it.

WINE did it for Win32 and Mono did it for .NET, so Java *will* be open source some day anyway. Sun needs to get at least J2SE out there before .NET runs on every electronic device available.

Now that Sharp's Zaurus has dropped Java, .NET is looking like the only alternative for managed coding on handheld platforms. (Cellphones are not yet good PDAs, ok?)

SLM

Bad idea

[cpuh0g]

Why do people just assume that saying the magic words "open source" will automatically improve a particular piece of software? Rarely do the most vocal proponents of "open sourcing" something actually get involved and start contributing to the codebase. It's all about religion for the most part, unfortunately.

I still fail to see the benefits of "open sourcing" Java. How will it be improved? It's not as if the engineers at Sun are stupid and don't know how to engineer enterprise software. Don't you think Sun has heard that same complaint from some major league/big $$$$$ customers and done everything they could to improve said performance?

Even if they *do* open it up, Im sure the slashdot community will still hate them because they don't use a GPL variant license. Its a lose-lose situation for Sun, I don't get why they would even consider it. Is there a business case that will generate a 9-figure revenue jump from giving away the source for Java? I don't see it, but Im sure someone around here will happily clue me in.

Re:Bad idea

[goldsounds]

Rarely do the most vocal proponents of "open sourcing" something actually get involved and start contributing to the codebase. It's all about religion for the most part, unfortunately.

Rarely do the most vocal critics of "open sourcing" something actually understand the rationale behind free software, which is NOT to have the best code, NOT to have the most secure code, NOT to ship a product the fastest, NOT to contribute to the code, NOT to get something for free, or even to become the "commodity" implementation of a specification.

The best reason to "open source" something is purely and simply the freedom to access the code behind the software you are running; the freedom to change, or port to another platform, the software that you purchased or downloaded. This is the original philosophy of the Free Software Foundation, and the GNU project, who were collectively the inspiration for the "open source" movement.

So if you're wondering why anything less than a GPL license is unsatisfactory to the hairy, unwashed free-software factinista, why don't YOU look up the facts and get a clue about the software freedoms that may, one day, mean that your descendants can read e-books, watch movies and examine the collected creative output of humanity unencumbered by the imposed obscurity of closed-source software, DRM and other impositions on our freedom.

Yes, this is about religion. It's about an idealogical divide between people who would rather have free-as-in-beer convenient software, rather than free-as-in-freedom software that preserves your rights. Frankly, your arrogant pragmatism nauseates me.

Re:Bad idea

[dubl-u]

You can't write Ruby on Rails in Java? Oh. Best tell these guys: https://trails.dev.java.net/ Enterprise developers deal with inadequacies in Java by doing what everyone else does: Writing their way around it. This is where struts, hibernate, ant, junit came from.

I'm sure the Trails guys will be the first to admit that Trails (like peers RIFE, Sails, and Stripes) is not as simple as Rails is. We Java developers are used to looking past the 30 lines of getter/setter boilerplate and other mysterious Java incantations. But that's not true for a lot of devs.

Both Struts and Hibernate prove my point in different ways. Struts gets around Java's limitations with reams of XML, a cure frequently worse than the disease. Hibernate is closer to the spirit of RoR, but that's not really Java either; they have to do heavy wizardry with bytecode manipulation and dynamic subclassing because Java doesn't give them the necessary power. And note that Hibernate came from outside Sun. Sun's attempt at the same the same thing was EJB, which has been deservedly, if belatedly, shot in the head.

As a senior technical person in a large enterprise I too seek the holy grail of letting business people enter their business processes into a pretty front-end and getting a fully featured website with full supply chain integration automatically generated.

Actually, that's not what I seek at all. I believe that software development is inherently complex; you can't ever automate the thinking away. What you can do is give the developer tools to aid their productivity, to make simple things easy and dangerous things hard. Let them focus their brainpower on the things that are actually worthy tasks.

For a data driven system, it's great. For a complex system integrating with multiple other systems, it's no easier to use.

Exactly. At the low end, it's easier; at the high end, it's the same (or perhaps a little worse). But what Sun is missing is the idea that development need not always be hard. Pehaps they'll learn that from Hibernate, which comes from an attitude very like RoR. If not, I hope they learn it from Rails.

And because it does so much magic under the skin, if that magic doesn't meet your needs, you struggle more to change it.

I agree completely. This is where I'm waiting to see how the RoR and Ruby communities mature. They've got the entry-level system taken care of, but can they provide a path to complex systems without the kinds of complexity cliffs that you see amateurs using Access or VB run smack into? I think that Ruby the language has horsepower that other easy-to-use tools don't, but we'll see if they can turn that potential into actual results.

Re:Bad idea

[killjoe]

"Tough shit. Want a free Java? Write one yourself. Want Sun to give you one? Explain to them in their terms why they should."

Because open source languages such as php, perl, python, ruby and C# are eating into your market share every day. People have already abandoned java for most web applications and it never even got off the ground on the desktop.

Finally Sun does not make any money off of java. They have said so themselves.

Re:If they do, it will all depend upon the license

[BrainInAJar]

"Open Source" + "Sun Microsystems" almost certainly = "CDDL"

Why Should Sun Do This?

[javacowboy]

Writing a fully compliant JVM takes a lot of time and a lot of effort, especially the class libraries. Sun spent years writing that code, and none of the JCP partners can be bothered re-writing it themselves.

IBM, BEA, Oracle, etc pay Sun to license their source code so they can release compliant JVMs.

So, it should be no suprised the the open *cough*IBM*cough* source community "demands" that Sun open source Java. Guess how much money a certain company would save getting free source code that they're paying to license now? In the same of "the open source community", they'd like nothing better than to get the #1 competitor's hard work for free so they stop having to pay them for it.

The Java spec is open for anybody the re-implement, the source code is viewable by all, and the JDK is a free download. Sun has stated that they won't stand in the way of Apache Harmony or any other open source project that aims for a full open source implementation of the JVM/JDK spec.

So what exactly is the problem?

Will it make the competion less desirable?

[IsThisBl**dyNameUniq]

Two major development platforms are .Net and Java. One is fairly Open Standard but not open source - and gets demands for Open Source. The other is not even open standard yet people accept this. Maybe the real issue is people can imagine a world where Java is totally open but don't ever expect .Net to be so don't bother discussing it (The wonderful Mono efforts aside)

Re:This would help

[XaXXon]

Please don't forget that utorrent requires Windows to run.

uTorrent is NOT self-contained. It requires the Windows API to run. This part of its footprint is not shown when you look at its memory usage, but that first 256MB of RAM that windows uses is the reason uTorrent looks so small.

They should not open source it.

[Anonymous Coward]

I am much happier with Sun's Java than most open source projects out there. It's very high quality. I know that I may offend some people, but I think it's higher quality than Linux (as an OS, not kernel). It's my opinion though. Sun managed to keep it standard is admirable. I think Sun also deserve to make money/own the property it created. Why not develop open source version of it instead of asking Sun to open source it? One answer I think is that Sun does not have enough resource to fix bug or bring out features quicker for something as large as Java. This is a good argument. I think it could be addressed differently than Open sourcing it. For example, manage the development better. Provide better incentive for users to submit bug fixes. Promote Java support service so that critical bugs a company needs to be fixed is fixed quicker (it's there, but maynot be promoted enough). I develop Java enough to know that it's very hard to have a perfect tool to test Java standard. For example, there's no clear spec for Gridbaglayout. What you see isn't enough to implement an exact replacement for what Java has. This is just a simple example to show that stardard is hard to make, hard to be changed quickly.

This brings another point about Java standard. I remember JSF has many bugs that it tooks months to years to be fixed because the standard was broken. I think Sun needs to be much quicker than now to address these issue. These big problem should be fixed in a couple of weeks, or couple of months (2). Most people don't wait for a technology for a year or two to adopt it. They use alternative tech. This is usually a one way street and Sun will loose those customers.

Re:This would help

[IntlHarvester]

Problem with this argument is that almost nobody runs Windows just to use uTorrent, while quite a lot of people run Java just for Azureus. The resources required for Windows are used by all applications (*including Java*), but most desktop users only infrequently use Java apps.

Re:Third-Party JVM

[LarsWestergren]

The other main problem is Checked Exceptions, which force a programmer to write "try{" before the body of every method and "} catch (Exception e) {}"

No, not EVERY method. Just methods that that can reasonably fail (for instance I/O related operations), and that doesn't "know" how to handle the problem themselves. This helps you create well defined APIs, which in my opinion is one major reason there are so many frameworks and open source projects for Java.

Although relatively useless (if not harmful), these checked exceptions lead to a minimum of 122 extra CPU cycles per method invocation.

Evidence of this? Besides, it has been said so many times, but appearently it has to be said again. Processing cycles keep getting cheaper. Programmer hours keep getting more expensive. Trading a few cycles for a feature that helps you create more stable and transparent code is sensible.

catch (Exception e) {}

That is just about the worst thing you can write. Ok, maybe catch(Throwable t) {} is worse. That the first editions of Bruce Eckels Thinking in Java books were littered with those is evidence he just doesn't get checked exceptions.

Re:What ? Eclipse has no issues with languages at

[JanneM]

I call bullshit - just because you don't know how to set up your machine properly doesn't mean java has language problems.

Great for you; I never got it to work properly (Ubuntu and SCIM/Anthy). I first had to add fonts to some java-specific list to get it to show CJK at all. When I run the app with Swedish locale it refuses to let me input Japanese (it does not listen to the SCIM server).

I'm sure I could get it working with enough effort - but after one frustrating evening I'm not going to bother. Java isn't alone out there; just about every Java app has good equivalents without the hassle (including the Kanji app I was trying to use). And I'm certainly not going to be using Java to develop anything knowing that potential users will have go through the same mess I do.

I should not have to "set up my machine properly" - most users do not have the technical skills to do so. I should be able to select "java" in the package manager (or rather, select the app I'm actually interested in) and it should all work - but it doesn't.

Re:This would help

[Anonymous Coward]

Programmer time is much more expensive than processor time these days. Therefore, many current programming languages are optimised to save programmer time first. C and C++ were designed in a time when processor cycles were extremely expensive, and therefore are optimised to save time at runtime instead.

As you have seen, java typically gets you results more quickly than C. In this case, since you simply took less time to get to your basic functionality, you could take more time to think about how to code more efficiently, and ended up actually writing faster code in the end.

However, java is not the only modern programming language out there. People have designed several new languages in the past decade. It seems reasonable to assume that some of those people deliberately set out to improve on java. Compared to such languages, java might appear to be very inefficient.

I'll leave it up to you to compare and decide. For instance, here's a comparison for web applications, done at JPL. (YMMV):

http://oodt.jpl.nasa.gov/better-web-app.mov

Re:If they do, it will all depend upon the license

[afd8856]

You may not realize, but "libraries are the new language". Seriousely, what good is a cool language if you have to reinvent the wheel everytime you want to have something like a report printer with print preview - a bit upset on python, with which I'm fighting right now to get something nice for a business app, I would really like something like jasper reports for python. Drawing on the DC with wx sucks for more then one form, and using reportlab to generate pdfs sucks as well. I've settled on doing a mono-platform hack, generating html with simpletal and calling its print preview dialog through ActiveX.

AMD64

[SIGBUS]

Maybe we'll finally see an AMD64 Java plugin for Firefox.

Re:This would help

[k8to]

Why are you arguing as if I stated that java uses more memory than C? Has downsides, sure, but the point is that Java is _the most memory hungry language in the entire world_.

You can go look at language shootouts showing example code and note how java always allocates the most memory. You can look at real world server applications (tomcat vs medus vs apache) or real world client applications (bittorrent vs rufus vs azureus). You will find that java is always using way more memory than the competition.

Java uses more memory than C.
Java uses more memory than C++.
Java uses more memory than Common Lisp.
Java uses more memory than Smalltalk.
Java uses more memory than Self.
Java uses more memory than Erlang.
Java uses more memory than Icon.
Java uses more memory than Pascal.
Java uses more memory than Simula.
Java uses more memory than Python.
Java uses more memory than BCPL.
Java uses more memory than Perl.
Java uses more memory than TCL.
Java uses more memory than Haskell.
Java uses more memory than Ocaml.
Java uses more memory than javascript.

There is _no_ common denomonator to these languages. Some have virtual machines as sophisticated as the jvm. Some have simple hand-hacked runtimes. Some are compiled. Some have features and dynamicism Java cannot hope to touch. Some are terse. Some are verbose. Some are forgotten and old. Some are quite new. Java uses more memory than every single one, and that is a major weakness of java in practical terms at this time.

rank amateurs

[mikaelhg]

why, oh, why java croaks on OutOfMemoryException when we have more than 8G of ram most of which is not being used

Because you are rank amateurs who are unable to read documents or use profiling tools such as jconsole or YourKit?

Re:This would help

[Decaff]

There is _no_ common denomonator to these languages. Some have virtual machines as sophisticated as the jvm. Some have simple hand-hacked runtimes. Some are compiled. Some have features and dynamicism Java cannot hope to touch. Some are terse. Some are verbose. Some are forgotten and old. Some are quite new. Java uses more memory than every single one, and that is a major weakness of java in practical terms at this time.

I know. This is a real weakness in Java. It would have been great if it was a far more memory efficient languages, because then it could have been used in a wide range of low-memory situations like embedded devices, PDAs or mobile phones....

Er - something wrong with this argument, perhaps?