Analysis Sun says open sourcing Java code will fragment and devalue the platform. Sun’s opponents say that under the current community process development is too slow. They’re both right, but the debate, which Scott McNealy regards as synthetic – an issue manufactured by hypocritical competitors – highlights what people really want from a technology. It’s an issue that finds Sun on the right side, but failing to convince skeptics. Read the rest of the analysis at TheRegister.
On Sun, Java and Open Source
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Finally! This article says it so right!!!
People just got to lay off Sun, though I grant it the people in charge really got to keep from saying wishy-washy statements. Instead of saying it will or wont, back and forth, say something smart like they will do what is best for developers. They got to be more smart in the PR position 😉 but other than that theyre doing what they should be doing.
At some point people in the community would dissagree and fork. and as Roberto J. Dohnert put it in his blog
http://rjdohnert.blogspot.com/2004/07/sun-ibm-should-quit-open-sour…
“..They make their own implementation and add features and sets to make their offering seem fresh. By the time the dominos fall and the smoke settles consumers need 7 different JVM’s just to view content and you make life a living hell…”
And as far as I am concerned, McNealy is correct on that one.
Schwartz seems to be playing the fool’s game. Bringing any kind of uncertainty over Java’s future makes .Net and C# look more attractive: at least developers know where the whole thing is going.
At the same time it shows Sun has lost its direction and focus.
Will people just cut it out with the whole “open source will fragmentate and kill Java” rubbishness?
There already are open source Java implementations!!
What do you think Kaffe and GCJ are?
So, where are the thousands of Kaffe and GCJ forks? While we’re at it, where are the thousands of Perl and Python forks? GNOME and KDE forks?
Furthermore, forks will not kill Java even if they appear, because:
1. They cannot use the name “Java” unless they’re officially certified. Java is a trademark.
2. Who would use an incompatible Java fork? Can you guarantee me that when someone forks Java, suddenly 50% of the community will develop for the incompatible fork?
As for point 2: if most people do develop for the incompatible fork, then that means the fork is better. It’s simply choosing the right tool for the job. If you cannot understand that then you are a zealot.
“At some point people in the community would dissagree and fork.”
And if the fork turns out to be better than the original then why is that a bad thing? Are you against innovation?
Actually, we already have open source JVMs!! Where are the forks? Where’s the incompatiblity?
I don’t think that Sun fears the OS community, I think it fears that if it were to go OS, Microsoft and or IBM would bastardize it so much, that it would die. Microsoft tried to do it before and I don’t see why they won’t try it again.
But since you already mentioned that there are OS versions of Java, why should Sun make it OS then? I fail to see any real benefit for Sun to make Java open. It’s already has open as it can be. “If it ain’t broke, don’t fix it”.
This is why the OS community will never succeed, it’s doomed to fail.
I have a solution for people who fear open-source Java forks:
Keep on using the official Sun VM.
That simple.
It’s ironic that most of the fork-FUD is coming from enterprise developers. If you’re an enterprise developer, chances are you only develop server-side, in which case you have absolute control over what’s running on your system; or else you develop in-house corporate desktop apps, in which case you have nearly absolute control over what’s on the system.
So just don’t use anything but the Sun VM. Just like you don’t use Blackdown or Kaffe right now. In other words, do absolutely nothing different; your world will not change.
I actually think the big reason so many Java developers oppose open-source is that they know it will draw many more (good) coders into the language. J2EE developers are already under pressure because corporations are realizing that the platform is overkill for a great many applications. Adding extra competition would only compound their problem.
(Compare the pathological hostility of Common Lispers to Schemers)
This is a classic example of “we had to kill the patient in order to save it.”
no way…that argument doesn’t hold water. MS was able to bastardize Java because it was the ONLY one to do so, and the ignorant users/developers supported them.
if many groups had the chance to modify Java, I doubt if they would act alone, and if their acting alone would be more than a mere drop in the bucket. most likely if they will modify (or “fork”) Java, they will do so collectively. and that modified Java will become the standard, if the plain Java sucks hard.
again, yes, Sun can simply open source Java but still control the trademark and what goes into standards. similar to what Linus is doing with the Linux kernel. where are the Linux forks? have you seen any, or have you seen any that has seriously undermined the whole Linux movement? To think Linux has been open since 1991.
the purpose of an open source Java is to provide a definitive reference implementation of Java. so it still makes sense to have one.
@Adam
Schwartz seems to be playing the fool’s game. Bringing any kind of uncertainty over Java’s future makes .Net and C# look more attractive: at least developers know where the whole thing is going.
Ahem. .Net and C# are already “forked”. (Mono). For the time being, Microsoft has actually been helping.
@Abraham
Microsoft and or IBM would bastardize it so much, that it would die.
Re: IBM –
What, just like Eclipse is now killing Java on the desktop?
Microsoft tried to do it before and I don’t see why they won’t try it again.
Re: Microsoft –
a) MS already tried to do it again: C#.
b) If Sun offered Java under a GPL-ish license, any incompatible changes that MS made would legally have to be made visible to the world. Would MS want that? Hell no. Also, any incompatible changes could be incorporated into the main VM.
Finally, Sun still has the option of still requiring any modified code to be passed through certification before it can call itself Java.
In short, nothing to worry about except 1) more eyes for bugfixes and improvements, and 2) more developers.
An all-Java-based GNOME desktop would be another small bonus.
>Sun can simply open source Java but still control the trademark and what goes into standards. similar to what Linus is doing with the Linux kernel. where are the Linux forks? have you seen any, or have you seen any that has seriously undermined the whole Linux movement? To think Linux has been open since 1991.
Well said! This is a very good analogy.
People don’t realize that a truly significant fork (one that really has the momentum and mindshare to fragment a market) takes a *lot* of time and energy. Forks just don’t spring up overnight.
The java specification is defined by a standards body using the java community process. This makes java more open than any other language.
There is no ‘open source programming language’. A language specification is always decided upon by either a standards body (e.g. C, Java), by a company (e.g. C#) or benevolent dictator(s) (e.g. Python)
The java specification is defined by a standards body using the java community process. This makes java more open than any other language.
The JCP is in practice a slow-moving body of corporations. Not that that’s necessarily a bad thing.
There is no ‘open source programming language’. A language specification is always decided upon by either a standards body (e.g. C, Java), by a company (e.g. C#) or benevolent dictator(s) (e.g. Python)
True, and irrelevant. The question is whether the Java runtime is open-sourced, rather than a trade secret as it now is.
I already anticipate your witty retort: “go and make your own JVM”, you say. Well, we already have a few. The problem isn’t the JVM per se but the core libraries.
Many of the core libraries are trade secrets[1] – and I mean the specs, not the code. One condition of viewing the specs is that you agree not to put out a partial implementation. This means, essentially, that one cannot see the spec and work on the OSS JVM.
Open indeed.
[1] See for example http://jcp.org/aboutJava/communityprocess/JSPA2.pdf and http://jcp.org/aboutJava/communityprocess/final/jsr129/j2me_pb-1_0-…
I don’t see why people keep on claiming that
Open Sourcing Java standard libs will cause evil forks;
There is already a comment here that states it correctly
(no one needs to use a fork,…), but still, here
is the link to my Open Source Java FAQ which (I think)
shoult counter all possible arguments:
http://www.jroller.com/page/murphee/20040426
Otherwise, what problems do you think the LSB was designed to solve? Also, the endless forking of linux into multiple different distros is a huge barrier for the ambition of getting it onto the desktop, considering how difficult it makes it to just bung an .exe on a CD or website and have it install everywhere.
The open source “javas” are not presently much of a problem because they are very far behind. Nonetheless, they currently fail the java compatibility tests, and this is why many a java program will not run on them.
C++ is also a good example of the dangers of industry forking. Back in the 80s, you had a dozen different companies producing their own special versions of C++, with various “value-added” features. This is how C++ became the impossibly complex beast it is today. Considering java is an industrial response to an interesting but fundamentally flawed (because of that complexity) technology, and tries to be more simple, I would not like to see java got he way of lisp, c++, and any number of other languages and platforms in endlessly accreting complexity and vendor “features” rather than being centrally stewarded.
Open Source fundamentalists have no moral right to the java platform. Their arguments that open sourcing it would somehow lead to some heavenly world and a much better java ring extremely hollow, given that if these claims of incredibly improved development speed were true, they’d have created their own open source and entirely compatible java implementation by now, of similar quality. But gosh, they haven’t! So who knows in what sense exactly open source development is “superior”. Hint to fanboys: drop the stupid ESR-style “open source is necessarily and qualititively *better* technologically line (by now so thoroughly debunked that proponents just look like laughable fundie zealots divorced from reality), and stick with a more respectable RMS-style line that open source is perhaps morally superior, in your opinion. But don’t turn this into some sort of mission to harass and blackmail companies as much as you can to release code for you. Just STFU and write some bloody code if it bothers you so much, and stop whinging for the labour of others like a bunch of ungrateful zealoty theives.
In my company open source is so thoroughly discredited by the actions of the zealots typified by the ones on this site that there’s little chance we’ll give it a look inside the next twenty years.
Oh, your whole premise is tosh. Yes, there are 1 billion Linux distros out there. How many are commonly deployed in business? 1, maybe 2. Want to write a Linux app? Compile it to Red Hat. If you have time, maybe SuSE. That’s all that’s required for commercial software.
Second, Sun likes to pretend the problem is with “forked” Linux kernels. In reality, however, incompatibilities between distributions have much more to do with the zillions of libraries and so forth that are packaged with it, or even just different directory structures. (Sun’s Linux Java VM, BTW, does not install itself where the LSB says it should. /usr/java/? WTF?). A Linux distro is a far more complex beast than a Java runtime. Incompatibilities are to be expected between distributions, which is why business has standardized on two of them.
The C++ analogy hardly holds water, since the incompatible compilers you’re speaking of were all proprietary.
Besides, languages that strongly depend on open-source implementations tend to be much more centrally stewarded than languages that do not. Does John McCarthy exercise the same control over LISP that Larry Wall does over Perl? What about Bjarne Stroustrup vs. Guido van Rossum?
Just STFU and write some bloody code if it bothers you so much, and stop whinging for the labour of others like a bunch of ungrateful zealoty theives.
Oh, calm down. Nobody in this thread has said “Sun must open-source Java or they’re evil”. All that’s happened is people have popped up to refute the fear-of-forking-FUD.
Sun can do what they want with the platform. But fear of forking is a cruddy justification for not open-sourcing. (Sort of like WORA is a cruddy justification for managed code execution – security and programmer productivity are much more important IMHO).
And by the way, it’s difficult to “write some bloody code” for specs that you cannot see unless you promise not to code for them. We in the free/open software movement care a lot about intellectual property – at least those of us who write the code do – and do not take such agreements lightly. Work on Free Java implementations continues apace, working around the obstacles placed up against it by the “open” Java standard.
Open Source fundamentalists have no moral right to the java platform.
I agree. Eric Raymond in particular should be ashamed of the crybaby demands he’s been making. Stallman, allegedly more of a zealot, has had a much better message: if you care about Free Software, don’t use Java yet; he hasn’t made any demands at all of Sun.
In my company open source is so thoroughly discredited by the actions of the zealots typified by the ones on this site that there’s little chance we’ll give it a look inside the next twenty years.
I am not confident in your management if they can be swayed by OSNews postings.
is well.. the big picture. if sun doesn’t open source java now, they stand to let C# and friends walk all over it on the linux desktop. the #1 reason java is shunned in the community for desktop applications it is requires “an annoying download from sun’s website” because distros are not legally allowed to distribute the jre.
if sun doesn’t do it now, and lets C# get enough momentum, they won’t be able to stop once longhorn (which is basically .NET++) comes out.
Seriously, if an specific implementation begins to get a bad rap because it messes up in many ways, people will just ditch it.
Look at the kernel for example, or another example, c++ compilers. Unless you mess things up terribly, at which point no one should be using your implementation, things will be ok. I think Java would gain more from being open source than the supposed losses of such a move. its too big, and you would have a lot of work to understand the codebase and start making changes to it, as long as you did things well, no problem.
I don’t think the “fear of forking” is just FUD at all. You have to remember that the java platform is jealously coveted by many very large organisations and groups of people with the will and capability to pull it in their own special directions. IBM want it open sourced so they can do a microsoft and create their own special java that has special hooks into weblogic and uses balls like SWT. Microsoft would love nothing more than to be able to take java and say “hey high end vendors, we have our own value-added java too”. Ditto many other high end vendors: it is difficult enough to keep them in the process as it is.
Point is, java has *already* suffered forking by other vendors with a more restrictive license, Why will opening it up make this less likely?
Oh, that’s right: because in open source java, every single extension made by any vendor would have to be added to the core platform to retain compatibility. I’m sorry, that’s not feasible, unless you want a disconnected tower of babel. For java to be open sourced, there would have to be a mechanism to stop vendors spreading their own special javas (those extensions being open source or not makes no difference if they are regarded as incompatible with the aims of the core java platform).
The reason, of course, that this does not happen with python or perl is that frankly those scripting languages are not exactly coveted by different multinational corporations with their own particular ambitions, aims, and desire to individuate themselves from the competition. also, they have grown up inside the open source movement anyway, so the open source movement itself creating a fork is less of a threat as it is with java. Still, if python and perl were central to a multi-billion dollar industry at the high end “enterprise” that MS wants to move into and IBM/Oracle/BEA et al defend, it’d be forked in no time (unless, as you seem to be suggesting, the details of every single fork and special variant are incorporated into the core platform, but you know, that would end up being a complete mess).
The simple fact is that java has been plenty successful, on the serverside and for in-house swift development projects, for the task it was designed to do (essentially being a practical, better cobl replacement). SO why the hell change the license? Because some open source folks want to use it in gnome? Bleh, these are the people that think java is unsuccessful because it isn’t used much in client side applications (ie ignoramuses). Why change something that works rather well as is? Beats me.
It all comes down to the covetousness of IBM and the open source movement, and their philosphical rejection of working inside the JCP on the core platform, to which anybody can contribute changes and bring the platform forward (Can you do this in .NET? No). Java is already open enough to anybody while still being able to retain its integrity.
Point is, java has *already* suffered forking by other vendors with a more restrictive license, Why will opening it up make this less likely?
You’re right, it has. And not only MS’s fork. An open-source license probably wouldn’t make forking less likely (although it would scare off MS), but it would make forks more repairable.
But more importantly, Sun can prevent any modified code from calling itself Java until it passes the certification. Don’t want to risk incompatibility? Don’t use something that’s not called Java.
Up til now the most significant Linux kernel fork (and I’m talking real forks here, not the penny-ante shit that J. Schwartz has been whining about) was the NSA’s SELinux. Although it was technically superior from a security standpoint, it did not see real adoption until parts of it were integrated into the core Linux branch.
Oh, that’s right: because in open source java, every single extension made by any vendor would have to be added to the core platform to retain compatibility. I’m sorry, that’s not feasible, unless you want a disconnected tower of babel. For java to be open sourced, there would have to be a mechanism to stop vendors spreading their own special javas (those extensions being open source or not makes no difference if they are regarded as incompatible with the aims of the core java platform).
For the core, the mechanism would be trademark law.
Besides which, there are currently a whole lot of un-standard non-core extensions floating around. What’s the procedure for dealing with them now? Mostly it seems to be running around and screaming that the sky is falling because IBM came up with a better gui library.
The simple fact is that java has been plenty successful, on the serverside and for in-house swift development projects, for the task it was designed to do (essentially being a practical, better cobol replacement).
Java was not designed for this. It was designed to run distributed client-side applications. J2EE came later. (And most of that spec, by the way, was written by IBM).
Bleh, these are the people that think java is unsuccessful because it isn’t used much in client side applications (ie ignoramuses). Why change something that works rather well as is? Beats me.
This is a pretty typical Java developer attitude: “Who cares about the small fries? My COBOL^H^H^H^H^H J2EE apps are going just fine!”
Many of today’s small apps will be tomorrow’s big apps. Slagging off the “trivial” stuff is a sure way to choke off the future growth of the platform.
As I’ve said, what Sun does with their code is their own business. But the future of the Java platform depends on attracting – as S. Ballmer would put – “Developers! Developers! Developers!” etc. Some of these are necessarily going to include ignoramuses who only care about the desktop.
Java is a big language. It can more than accomodate client-side morons alongside us server-side geniuses. My big problem with the JCP is that most participants don’t seem to share that sentiment.
Do you consider Linux Kernel 2.6 a fork of 2.4?
I don’t think so, even 2.6 would exhibit different behavior from 2.4 in some aspect.
Has Mozilla been forked like mad? Or any other open source projects with forking so seriously that it is completely chaos? Maybe some of them, but for most it’s not that big of an issue so far (unless it involves licensing changes).
The Linux forking is clearly exaggerated. I downloaded Mozilla binary from official site and it never fails to run on different Linux distros I have tried. There is some common base of dependency you can build generic application to run on, I am sure.
Now it is a different question if Sun could lose control of Java; since open source, after all, is also a form of competition. If IBM, BEA or other companies can crank out better JVMs or extend its feature (alone, or together), Sun will need to keep up, or risk losing relevance to Java (ok, maybe it’s ‘a fork of Java’ that cannot be called Java).
Linus Torvalds does not derive profit from the Kernel, nor the Mozilla developers stand to gain big cash from it – that cannot be spoken of Sun in their position of Java – in its control of the spec and other licensing agreement (currently you cannot bundle a Sun JVM with a Linux distro).
Moreover, the experience of Mozilla shows it could be a few years of re-write before a new, reliable architecture become mature and surpass the old; who knows how keen developers would want to engineer a faster Java – and take how long?
In the end, I think an open source implementation of Java will be better for developers and FOSS community; however, the risk and uncertainty involved in migrating to open sourcing Java is not to be underestimated. Perhaps, Mono will show whether OSS is really a critical factor for drawing developers’ interest and mindshare.
Linus Torvalds does not derive profit from the Kernel, nor the Mozilla developers stand to gain big cash from it
I don’t know about the Mozilla developers (I certainly hope the guy in charge of XUL documentation hasn’t seen a penny), but Linus makes a decent salary.
that cannot be spoken of Sun in their position of Java – in its control of the spec and other licensing agreement (currently you cannot bundle a Sun JVM with a Linux distro).
Actually, it can be spoken of Sun. By most reports the profits of their Java division are quite modest. It is their most recognized and respected brand, but so far they’ve had limited success in monetizing it. They’re not losing money, but they aren’t raking it in, either.
Which, from the perspective of a Sun investor, would be an argument in favor of open-sourcing Java: stop throwing cash at the damn thing and move onto something with better return on investment.
(currently you cannot bundle a Sun JVM with a Linux distro)
You can, but only in the shrinkwrapped ones where the distributor has a license with Sun (e.g. SuSE, Xandros, etc.) Doesn’t work for the free/download versions (Slackware used to have a JVM, but they had to give it up with 10.0).
Moreover, the experience of Mozilla shows it could be a few years of re-write before a new, reliable architecture become mature and surpass the old; who knows how keen developers would want to engineer a faster Java – and take how long?
Oh Jesus. Mozilla’s problem was they thought open-source was magic pixie dust and that it exempted from every known rule of software engineering: don’t rewrite from the ground up, don’t write junk you probably won’t need, etc. If Sun should decide to go forward with it, I hope they learn from Mozilla’s example
> I don’t know about the Mozilla developers (I certainly hope
> the guy in charge of XUL documentation hasn’t seen a penny),
> but Linus makes a decent salary.
The point is he’s not selling any license of it.
In the past Linus refused to join the board of directors of a Linux related business (which he bears no operation responsibility, just want to bear his name as advertising I suppose); he believed it would compromise his neutrality (or community’s perception of it). I believe that kind of character would help in leading a big community project like the Kernel.
> Actually, it can be spoken of Sun. By most reports the
> profits of their Java division are quite modest. It is
> their most recognized and respected brand, but so far
> they’ve had limited success in monetizing it. They’re not
> losing money, but they aren’t raking it in, either.
Interesting, but can OSS Java provide more business? That’s a question need to be answered to investors.
> You can, but only in the shrinkwrapped ones where the
> distributor has a license with Sun
Not all Linux distro then =P .
Certainly not for free distro w/ ISO.
> If Sun should decide to go forward with it, I hope they
> learn from Mozilla’s example
It is the question, can Sun lead such a project without repeating Mozilla’s mistake? Sun will need to have a vision already on what direction Java is going; um.. do we see that yet? Does Sun want to put Java on desktop? If not, then what’s Sun’s vision – which can take advantage of the open source development process?
>Moreover, the experience of Mozilla shows it could be a few years of re-write before a new, reliable architecture become mature and surpass the old; who knows how keen developers would want to engineer a faster Java – and take how long?
Well, the initial Mozilla codebase was so bad, they had no choice but to start again from scratch. (Anyone remember the buggy-as-hell Netscape 4?) I don’t think the Mozilla scenario applies here. I’ve worked with Java extensively for over five years and I’ve never experienced any problems with instability or crashes due to the VM.
>In the end, I think an open source implementation of Java will be better for developers and FOSS community; however, the risk and uncertainty involved in migrating to open sourcing Java is not to be underestimated.
I think you’re right here. As far as I’m concerned Java is open and free enough already. But I would be happy if Sun just relaxed the terms for distributing the JRE. Mainly the ability for vendors to distribute their a customized, stripped-down JRE with their application. (I’ve posted about this before, so I won’t bore you people with this again.)
>Perhaps, Mono will show whether OSS is really a critical factor for drawing developers’ interest and mindshare.
It usually takes several years for a technology to mature and attract enough interest, Mono will not be any different. Despite all the hype and the fact that it has been available for a while already, .NET is not being used widely yet for desktop applications. Until the .NET runtime becomes a standard Windows component, this will not happen either. As with Java, installing a separate runtime is just a huge stumbling block for users.
As a side note, can IKVM run Java apps on Mono?
IKVM
http//www.ikvm.net/
“As with Java, installing a separate runtime is just a huge stumbling block for users.”
i do not agree with that, if you provide a decent application that users like, they can install even the beta versions of Java Runtime not matter the size is, Check Azureus, it is just a bittorrent client, and because it is a good piece of work, millions of people downloaded it and the latest JRE in one week. (To be correct, only this week 1.6 million downloads. )
http://sourceforge.net/projects/azureus/
As a side note, can IKVM run Java apps on Mono?
Some of them. I’m not sure if all the libraries are implemented.
Interesting, but can OSS Java provide more business? That’s a question need to be answered to investors.
A lot of Sun’s investors wish they would get out of Java altogether.
They make some money off it, not a lot. The big money is in Java application servers. Sun’s share of that market is small, while IBM and BEA’s are large.
The argument would be that they stop devoting lots of resources to Java and devote them to something more profitable. (The problem, of course, is what the more profitable thing would be.)
Actually, it can be an issue. Sure, with something like a Bittorrent client or Limewire, the download is fine. (Although this /usr/java thing is a real annoyance).
But on Linux at least, if you want to use Java for core functionality – OpenOffice or GNOME applications or whatever – it is an issue, since you can’t package it into the free/download distributions.
Downloading and installing the JRE remains an issue for ordinary users, not necessarily for the technical crowd here on OSNews or SourceForge.
I think Azureus is a great “flagship” product for proving that Java desktop application can be appealing and competitive with their native counterparts. (Thanks to SWT.) But still, like most SourceForge projects, Azureus is really an app for technical people.
That all said, LZMA compression can crunch down the JVM to a reasonable size. The default Windows distribution of my JPluck[1] app is 3.4MB in size, while the version with the JRE included is 12.2Mb.
The download statistics for JPluck show that there is a sizable number of people who download the JRE version. For some versions, they even constitute a third of the total number of Windows downloads. (I’m pretty sure that there are plenty of people who were put off by the Java requirement and chose not to download it at all.)
[1] http://jpluck.sf.net/
off topic,
JPluck seems like a cool application with a decent GUI. i will give a try. JGoodies look and feel i presume?
Yes, JPluck uses JGoodies L&Fs and the Forms library. I still regret that I didn’t choose SWT, though.
[off topic]
well, actually just opposite, i like swing better. i find SWT a) harder to program(or a differnet approach, lets say) b)not so good in terms of speed and shape in OS’es other than windows. but this is personal preference. Both are fine and should exist.
But then we wouldn’t have all the “Open Source or Open Sores” jokes anymore. So, I don’t know what to say about it.
Well, I agree that SWT is slightly harder to work with and that on non-Windows platforms the implementation is still not up to scratch. (Though I must say that I haven’t seen Eclipse 3 on Mac OS X yet.)
But no matter which way you cut it, a native UI is always going to feel more responsive and look better, or at least more familiar, to the user than a Swing UI. I’ve heard plenty of users complaining about Swing apps looking “funny” or feeling slightly “off”. It always amazes me that some of the most rabid Swing advocates just sweep these concerns under the carpet. Aside from the look-and-feel, the huge memory requirements and long startup times still make Swing apps stick out like a sore thumb.
At any rate, the choice between SWT and Swing isn’t clear cut. It all comes down to your project’s specific requirements. Swing is used a lot for creating specialized applications in vertical markets. (Just check out Swing Sightings for numerous examples.)
Even so, I really believe that native UIs in apps like Azureus will do more to advance the idea of “Java on the desktop” than Swing ever will.
> A lot of Sun’s investors wish they would get out of Java altogether.
If this is true, perhaps we could ask a few Sun’s investors to speak on this matter in public. This would have more teeth than some ‘great fame lobbyists’ from the OSS community.
Because of the fragmentation of Linux you can obviously see how it has been devalued, at least for Richard Stallman.
Seriously, out of the billions people are paying for Linux solutions how much of that do you think the FSF gets? Well, if Sun opened java it would become something more like Linus/OSDL or the FSF as far as its profit center is concerned.
That’s a bad business decision for Sun, a company that cares much more about profitability than technology, IMO.
> Well, if Sun opened java it would become something more like
> Linus/OSDL or the FSF as far as its profit center is
> concerned.
I don’t think Sun will become OSDL / FSF thru open sourcing Java; they are non-profit by nature. What will happen is if Sun can make itself like MySQL, JBoss or even IBM (if it has the level of expertise to do so) by customizing Java for its customers. Of course MySQL and JBoss are much smaller companies to begin with, and I am not sure if Sun really wants to put Java on Desktop…
It is only when outside companies innovate that forces Sun to update Java. SWT is a good example.
Left to their own devices, Sun’s version of Java would just stagnate — or the opposite, get infected with giant blobs of poorly designed class libraries (J2EE).
Because Sun is in such complete control, there is little to no innovation in Java.
I should point out Microsoft .NET has the same issue — because it is fully under the control of Microsoft it is a dead platform, not one that will expand other than by stealing and assimilating the works of others.
IBM has their mission understood very well — the computing base only grows when the common platform is owned commonly. Sun’s control of Java is not even good for Sun as the base is stagnating because of Sun’s iron grip.
Sun is a very stupid company when it comes to basic understanding of market dynamics. Most of their executives come across as highly ignorant, highly egotistical, and highly unqualified.
Ironically, one can only hope Sun will “see the light” before it is too late. Otherwise it will be “we had to kill the platform to save it”.