Java is an open specification. The libs are open specifications. Just because the FSF hasn't been able to finish an implementation doesn't mean it can't be done.
Precisely correct. If the Open Source community would listen to reason instead of RMS and ESR's politics, it would embrace Java where it is the right tool for the job (enterprise applications with J2EE) and then make a concerted effort to bring GNU ClassPath and Kaffe to completion, eliminating the remaining proprietary ties. Where RMS and ESR go wrong is that this really IS the same situation the GNU project faced before glibc, gcc, and Linux. The tone of RMS's article implies that currently available Free Software is always a suitable replacement for Java. This is dead wrong. There are no equivalents to J2EE anywhere, unless you only need to write a lightweight web application. (in which case you can use PHP or Python/Zope..)
"Many of the things that I can do in java would be very hard in any other language..."
You really need to get out more. But I won't waste more space here debating technical misperceptions, this is about freedom.
Where you and many others are wrong is that Java, via J2EE, provides a solution that no "pure OSS" solution currently can. Some of us write things more complicated than simple web apps. Enterprise applications that is. There is no equivalent to EJB in Python, for example, and the other tools roughly equivalent to Servlets, JSP, JNDI, etc. are immature as well. I use JBoss's J2EE and Java tools by the Apache Project, all 100% open source. All I need is a 1.4-equivalent JVM and matured GNU ClassPath. This is very much the same situation that RMS described regarding the early days of the GNU project -- some of the needed tools were not yet freed.
But there's no way you can make it free. Only Sun can do that, and your seeds aren't falling inside their walls
False. Any company with a clean-room JVM could help to make Java free at any moment. Even if they do not, Open Source JVM's and GNU ClassPath will get there eventually. (hopefully faster once people realize why Java is really needed)
So what jobs are you doing that make Java the best solution?
Enterprise applications. The Python / Perl / PHP world is currently unable to compete with J2EE for complicated, professional business applications. And note that I'm talking about more than just "huge websites." I'm talking about the core software that businesses rely on for daily operations. The P* tools are good for lightweight web applications, but that's about it. Python tools may have the potential to compete with J2EE someday, but they have a long ways to go. Zope + something roughly equivalent to EJB could perhaps work for smaller business apps.
I don't see how things have come out any better for Sun than if they had set Java loose.
It would have led to broad and rapid adoption by the Open Source community. How successful has MS been at unseating Apache? If Java had been completely free and open from day one, it would likely have enjoyed massive market penetration. Once a really solid standard is in place, it's hard to shake it..NET would have been the underdog, just like IIS. Now it's going to be an eventually open source "Java" trying to catch up to.NET's market momentum. Surely possible and a good goal, but it'll be harder now.
Quite frankly, the stupidest thing Sun did was force MS to give up Java. MS wanted to make Java ubiquitous by making it the standard platform for writing Windows apps.
It's not so much just what they did, but that they used monopolistic tactics to do it. MS was trying to completely take over the Java market, not just be an ethical player. In and of itself, there was nothing wrong with them writing their own implementation. (many others have done the same, including Open Source projects..) It was all the proprietary and undocumented extensions combined with their marketing power to sway the industry away from pure implementations without their extensions. This would have hurt us all, not just Sun.
How do we feel about MS's extensions to standards like Kerberos and HTML? Well, same thing.
First I'd like to express my opinion that Cringely is dead wrong. The war is far from over.
Linux needs to do something *groundbreaking* that Windows doesn't, that Microsoft can't suddenly copy, *and* that the public actually care about.
Absolutely correct.. I have three suggestions of things we can do that MS cannot duplicate:
1.) The Windows + Office duo is MS's cash cow. If you kill the cash cow, you quickly shut down the farm. The way to do this is OpenOffice, but OO.org needs a lot of help. Somehow, the Open Source community needs to figure out a way to fund OO.org and bring on a few dozen more full time coders to the project. Not forever.. just until it equals and begins to surpass MS Office. MS cannot give away Office on a large scale. It would kill their bottom line, they'd start reporting huge losses, and their stock price would plummet. And if the stock price plummets, those billions in the bank start getting used rapidly. For many people, the reason to run Windows is to run Office. Many large business workstations exist for the purpose of Office + web + internally developed apps. If you sweep the market for office suites, you get a lot of the market for Windows as well. They go hand in hand more than most people realize.
2.) Network booted workstations. We need to write tools that make it trivially easy to set up and administer a lab or office full of diskless, network-booted Linux workstations. The implications here for TCO are enormous for two reasons: One, you eliminate the most unreliable component of the workstation (cheap IDE disk). Two, the administrative costs fall dramatically because you only have to deal with the server. No more going around to 100 machines to install software updates, fix configurations, etc. Because of the licensing nature of proprietary software, this would be a difficult act to follow. Add Paladium and heavy-fisted DRM to the mix and it becomes near impossible! So even if MS started giving away Windows and Office, they'd still be losing the TCO war. And TCO is what most PHB's really care about..
3.) Business software. It's one thing to have a free OS and basic desktop software. It's quite another if businesses can rely entirely on free and easily customizable software for all their needs. Business software is typically not purchased from MS, although traditionally it requires a MS OS. So if that software is free and runs on a free OS instead, there's a much stronger incentive to ditch Windows than has ever before been presented. MS can't compete because this is software they don't themselves produce. Again, even if they give away Windows and Office, businesses still have to buy all the other software. The OSS community needs to fully embrace J2EE as the right tool for the job here. Then we need consultant-developers to take the software into the field, improve it, and interact with the user community. *hint hint* Anybody 'round these parts un/under-employed?
In general your points lack merit and an understanding of where the enterprise server market is today, but one point I will pick at specifically:
"..and J2EE is too little and too late"
This could not be further from the truth. J2EE is a truly excellent solution for developing enterprise applications and.NET is currently unable to really compete with it, from what I've seen. J2EE is massively popular and dominates its respective market the way Apache dominates web servers. In my opinion, J2EE needs to be fully embraced by the Open Source community for the purpose of developing free alternatives to high-end business software that hundreds of thousands of companies need (or would like to be able to afford!) Enterprise software is the final key to locking in the future of Open Source permanently. It's one thing to get a free OS and relational database. It's another when your whole business (or non-profit) can run on free, collaboratively developed, and easily customizable solutions. (Think of an "Apache" style project for all major business applications: accounting, ERP/CRM, communications, document management, etc.) For business purposes, the desktop is increasingly meaningless today -- it has largely become commoditized. (OS + office suite + web browser) Powerful, modular n-tier enterprise applications are the future.
..copy-protection software this summer that will for the first time give portable digital music players access to rented tunes from all-you-can-eat subscription services
Right.. this will definitely fly.. because you know, it's not like the market is already saturated with 100's of millions of DRM-free home, car, and portable mp3 players.
but keep in mind that most PC users think MicroSoft is GOOD.
Is that why half the people I fix computers for have asked me at one point, "When is this Linux thing going to be easy enough for me to use? I hate this buggy Microsoft stuff that gets all these viruses and spyware and constantly needs fixing." At which point I usually tell them, "wait about another year and check back with me." I've been saying that for about 3 years now. But it's getting pretty darn close to the point where I can start making recommendations for some people to consider switching to one of the user-friendly Linux distributions. (People who have practical needs and don't care about the latest games and gadgets. They just want a computer that works without error.)
IBM should just open source their VM under a different license and not call it Java.
They can't or else they already would have done so. IBM's JVM is based on Sun's code and thus they have no rights to open source it. What IBM (and others) can/should do is financially support the Kaffe and GNU ClassPath projects, which are clean-room Java implementations written by people who have never been tainted by exposure to Sun's "community source" licensed JDK code.
As a sidenote, even Sun is legally unable to open source most of "their" Java code, so we really can't bash them for saying no. Third party contributors hold rights to much of what is contained in the official Java distribution. Sun really screwed up by not licensing Java under the GPL from day one, in my opinion. I believe they would have held just as much sway over the evolution of the language -- both by holding the trademark and by being the trusted leaders.
My prediction: Java will be the next Unix. It's a close historic parallel. Even more sickeningly ironic: perhaps Microsoft's.NET is as to Java as Windows was to Unix.
Or, of course, the other option is we (the open source community) invent something totally new that is neither Java nor.NET. Some would suggest Python, but I don't believe it has the potential for enterprise-class solutions. (yes, it's a great language otherwise.. just in a different league)
From what I've seen, Magnatune is by far the closest yet to doing things "right" when it comes to indie music. The total crap stuff is filtered and you're left with unknown but still professional musicians. (ie. no high-school garage band wannabe musicians) DRM-free + FLAC is the only reasonable choice for quality and compatibility. (yes, I'm an audiophile and quite picky.. If I pay for music, I want full CD quality, not 90%) And as for previewing albums, their streamed 128kbps MP3 is fine.
That being said, one big thing is missing: popularity. It's one thing to find music that I myself like, but there's still a missing link of how to get the best artists onto national tours where they can make a career-level income. So the question becomes this: will the indie music pay download sites enable sufficient popularity to make this possible? I have my doubts.
Radio + bootleg P2P is a powerful marketing machine for major label music, even if the RIAA has not yet acknowledged this. (which is the real reason you shouldn't share RIAA music!!) Always remember that the real money is made in concert. Whether somebody buys a CD or downloads via P2P, the only thing that matters is popularity. Popularity breeds concert ticket sales. So how can indie music compete with this?
It's a catch-22. On one hand, the musicians want income now -- so they try selling the music itself. On the other hand, they want to eventually become popular enough to go on tour and make a real living doing shows larger than bars and clubs. I'm not saying there's one answer for everyone, but I'm not aware of any bands yet who have fully embraced the marketing potential of unrestricted P2P and "Here's some free CD's.. Please make all your friends copies" Don't get me wrong, this would only work for top-notch bands in mainstream genres, but it seems the only way to compete with the major label marketing machine. For the less popular styles or for non-career musicians with no desire to tour, sites like Magnatune are probably the way to go. Hopefully an increase in independent (and online) radio popularity will help as well.
it's just simple market economics. i'd say driver availability pretty much follows relative market share.
Except that the "relative market share" here was not determined by market economics. Monopoly is still the driving force even if hardware manufacturers are only doing what economically "makes sense" to them. Incidentally, there is virtually no cost in releasing complete hardware documentation so that alternative drivers can be written. And yet most hardware companies still do not do this -- many still refuse, others release partial documentation that only allows basic functionality. This is the reason why many Linux/BSD drivers are so behind. Of course, in some cases, it would be ideal for the hardware manufacturers to help out with open driver development because so much is at stake for them. Video drivers come to mind because the volume of sales is significant. (ie. just about everybody buys either ATI or NVidia chips)
There are some software titles that just -need- to be ported to linux, do to lack of OSS alternatives.
Wrong. We don't "just need" any proprietary software to be ported to Linux. We do need to get behind the projects that are developing OSS alternatives and support them both community-style and financially. We also need to gather support of the business community, focusing on software that will save them money. ex.) "You spend $10,000/year on Macromedia tools? Support our project and you can drop that expense within 2 years."
I personally would gladly donate $100 to a professionally run project implementing SVG solutions so we can ditch Flash once and for all. I would donate more if it would give me a vote in future feature development. If we can buy Blender in a few month's time, we can surely pool enough resources to do this.
Exactly! Don't buy RIAA music. Download your sh** online
What you describe is only half-right in terms of solving the problem of a corrupt entertainment industry. The correct solution is this: Don't buy RIAA music but support independent / local / non-RIAA artists. That's right -- don't even share RIAA crap. Doing so only makes it more popular - and thus keeps people buying CD's and merchandise, watching MTV, and going to RIAA-artist concerts. And. incidentally, Hollywood is another good boycott target. Don't want DRM-laden HD-DVD's and HDTV components? Stop buying today's DVDs and going to every movie that hits the theaters! Cancel your ridiculous cable/satellite premium package! These people can do evil things only because YOU enable them with your dollars.
Look to software as an example. The answer to Microsoft's monopoly is not warez sites; it's Open Source. And it's working.
When alternatives exist to fight corruption, the legal one should be chosen first--not necessarily because the law is just, but because it's the easiest path. Unjust laws can be changed far more easily after monopolists have lost the reins.
I am with you on this one. Sun opens up Java and IBM does the same with their Lotus suite? A gesture is met with a gesture and we all win. It will never happen.
IBM is not just saying, "Hey Sun, open up Java now!" as ESR did. They're saying, "We want to collaborate on an Open Source Java." There's a big difference. Read Sun's response to the ESR paper. Here's the relevant quote: "Sun would support an Open Source version of Java, but it need a lot of money and time to do so. You can't just flick a switch." In other words, Sun doesn't own all the code to Java so they can't just change the license. Other people have contributed code under the SCSL, which is not compatible with GPL. And Sun, being financially stretched as it is, does not have the resources to re-write all the Java code they don't control.
What IBM is saying is, "Yes, there is a significant amount of work to be done in bringing about a truely Open Source Java (ie. re-writing parts), but we're willing to help out and in fact bring most of the development labor" What Sun can do, if they're smart, is to begin releasing code that they control under a dual license just as with OpenOffice. And after IBM and others fill in the gaps, Sun can certify the code as meeting the spec.
Incidentally your example with Lotus suite makes little sense. Java is as much a piece of infrastructure software technology as Linux. Because of that, it needs to be truly open for maximum acceptance and innovation. Infrastructure is where Open Source has proved the strongest thus far. (OS's, libraries, compilers, server software, etc.) Although I do think IBM should open source Lotus someday too.
And what if Sun refuses to help with the Open Source Java effort? IBM should just go off on their own. They have the resources. They could bring together all the existing fledgling Open Source Java projects and unify them. This, of course, would not be beneficial to Sun and would likely lead to their demise, simply through diminished mindshare. It's really Open Source or bust..NET is looming on the horizon, gaining acceptance fairly quickly. Sun needs all the help they can get.
BS. The article gives several strong motivations, all of which I have personally run into as reasons NOT to use Java at this time. Java will not really take off until it is included with every Linux distro and can be fully embraced by the Open Source community. And it desperately needs more innovation, which OS community support would quickly provide. Sounds like you didn't RTFA. One thing the article didn't mention is that Sun's Java implementation is probably the worst in the industry.
It will be interesting to see whether the Linux world can do a better job than Microsoft at providing security out of the box in order to protect these non-techie users from their own ignorance. Will Linux distros be able to stand up to countless virii targeted at Linux?
The design of Linux/BSD and the software culture of Open Source make for a quite unfertile environment for viruses and worms. Sure, it's possible that somebody could find a buffer overflow in a common userspace program like KMail and exploit it, but not everybody uses KMail.. not even a majority. (compare the majority IE/Outlook users). There's a bit more security paranoia as well. Feature bloat is kept to a minimum in areas that could potentially be exploitable with a discovered flaw. It also helps that, unlike with Windows, even clueless users of Linux/BSD don't use privledged accounts for daily use. Will we ever see workable malicious code? Possibly. But it won't be an internet-wide epidemic. There, however, is a case for further security-hardening Open Source desktops before they become more widespread.
By the way, "virii" is not a word. It's just viruses. (:
Let's face it. Linux isn't for just the uber-geek anymore. So logically, more systems are going to be hacked into when people with no security sense are managing systems... Don't blame the operating system. Blame everyone who thinks they're a competent sysadmin, but really aren't.
It's true, Linux is not just for geeks anymore. But because of that, we need pre-hardened distros (including ACLs, IDS, and stack protection) and automated security updates for systems run by idiots. The ultimate answer (educating people) is unfortunately not feasible. As much as possible, security needs to be idiot-friendly on every OS.
Intel is obligated to its shareholders to protect it's technology. Open source drivers could tip their cards to AMD or some newcomer could gain the upper hand. That is the REALITY of how the hardware business works.
BZzz.. wrong. You clearly have no understanding of how the hardware business works. #1.) It is common practice, when actually necessary, to reverse engineer a competitor's product to see how it works. Lack of source code for their drivers is not a hindrance. Hobbyists have in the past reverse engineered complex hardware in their basements in order to write open source drivers. Multi-million dollar engineering labs are more than capable. #2.) It takes too long to copy off somebody else's design. If you do, you'll be too late to market. #3.) Intel is obligated to its shareholders to make money, not "protect its technology." Sure, there are cases and ways in which it sometimes can/should, but this is not an example of one of them.
Do you reall think there are people who say "let's not do this open source project in Java because its not included with XYZLinux 9.3"
Yes. This sentiment runs fairly strong in the Open Source community. Many would like to use Java but do not because their software will then not be picked up by the major distros. This is a perfectly legitimate issue.
So are you referring to desktop Linux? That's such a small marketshare
It won't be a small 'marketshare' for much longer, and secondly, the use of Java for desktop applications is finally showing potential. Free Software on the desktop needs Java and Java needs Free Software on the desktop. Enterprise Java is going quite strong, but it's not enough to support Sun.
and again it's not like most Linux desktop users don't know where to get a JVM. So it's hard to see how there would be JVMs on Linux machines that currently do not have JVMs.
Wrong. Most of today's new desktop Linux users are clueless folk, not developers. They use whatever comes with their distro. If JVM is not included, they go without. Look at it this way: why isn't Java more popular on the Windows desktop? Because Microsoft doesn't include it by default! (of course this is for anti-competitive reasons, not licensing)
Sun should also consider this risk: If a third party Open Source project implements a compliant JVM and JDK, it will quickly become the dominant Java implementation on all Linux/BSD desktops and perhaps even beyond. (especially if it becomes faster than Sun's JVM with refinement) And it won't be called Java -- it'll just be compatible with the spec. So much for mindshare. This is the inevitable outcome in the next 5 years if Sun does not GPL as much as possible of their own implementation. Here's to hoping they do so we don't have to needlessly reinvent the wheel.
Personally, I don't think something like Java will gain any benefits from following the route ESR proposes. By setting the Java source code free will fragment it more than ever. And for an industry that needs to hold off M$ as much as possible, I think this would be a bad move.
The benefit to Sun of GPL'ing their Java implementation would be expansion of their market influence. Right now, there aren't very many open source Java apps (comparitively speaking). This would change rapidly if a complete JVM/JDK could be included legally with every Linux/BSD distribution. Complete adoption of Java by the Open Source community would mean a sharp rise in the popularity of the language and this would help Sun tremendously.
Keep in mind that if Sun GPL'ed their Java implementations, it would not mean a true loss of control. They would still own the Java and related trademarks. So even if somebody forked Sun's GPL code, it couldn't be called Java. And, in like manner, Sun would still control the specifications defining what "Java" is -- they would still have the right to certify what is and is not "Java". In reality, the situation would be no different than today, where 3rd parties are welcome to write their own Java implementations using the open specification.
So in the end, both ESR and Phipps are each right on certain things. But Sun has no advantage in keeping their JVM/JDK sources under a license more restrictive than GPL. The other question perhaps, is whether something legally prevents Sun from changing the license -- 3rd party code, etc.
I personally see Zope as the Free Software answer to JBoss, and I think that it is pretty darn good answer as well. As an application server Zope is a fairly amazing piece of work. However, it won't run your applications designed to run in a Java application server
Both are of themselves just tools, but Zope is a completely different type of application server than JBoss. Zope is designed for writing web applications, not enterprise business applications. You would never write a CRM / ERP / Accounting application using Zope because it doesn't even have the beginnings of the requisite underlying functionality. Zope is good for portals, simple e-commerce apps, content management system, and the like. It is not a 'competitor' of J2EE/JBoss or.Net even though there is a small amount of overlapping functionality.
Somebody mod my post back down. I had received incorrect information regarding this issue. (Or was this a former Sun licensing policy that no longer exists?) My apologies.
Slashdot Mirror
Java is an open specification. The libs are open specifications. Just because the FSF hasn't been able to finish an implementation doesn't mean it can't be done.
Precisely correct. If the Open Source community would listen to reason instead of RMS and ESR's politics, it would embrace Java where it is the right tool for the job (enterprise applications with J2EE) and then make a concerted effort to bring GNU ClassPath and Kaffe to completion, eliminating the remaining proprietary ties. Where RMS and ESR go wrong is that this really IS the same situation the GNU project faced before glibc, gcc, and Linux. The tone of RMS's article implies that currently available Free Software is always a suitable replacement for Java. This is dead wrong. There are no equivalents to J2EE anywhere, unless you only need to write a lightweight web application. (in which case you can use PHP or Python/Zope..)
"Many of the things that I can do in java would be very hard in any other language..."
You really need to get out more. But I won't waste more space here debating technical misperceptions, this is about freedom.
Where you and many others are wrong is that Java, via J2EE, provides a solution that no "pure OSS" solution currently can. Some of us write things more complicated than simple web apps. Enterprise applications that is. There is no equivalent to EJB in Python, for example, and the other tools roughly equivalent to Servlets, JSP, JNDI, etc. are immature as well. I use JBoss's J2EE and Java tools by the Apache Project, all 100% open source. All I need is a 1.4-equivalent JVM and matured GNU ClassPath. This is very much the same situation that RMS described regarding the early days of the GNU project -- some of the needed tools were not yet freed.
But there's no way you can make it free. Only Sun can do that, and your seeds aren't falling inside their walls
False. Any company with a clean-room JVM could help to make Java free at any moment. Even if they do not, Open Source JVM's and GNU ClassPath will get there eventually. (hopefully faster once people realize why Java is really needed)
So what jobs are you doing that make Java the best solution?
Enterprise applications. The Python / Perl / PHP world is currently unable to compete with J2EE for complicated, professional business applications. And note that I'm talking about more than just "huge websites." I'm talking about the core software that businesses rely on for daily operations. The P* tools are good for lightweight web applications, but that's about it. Python tools may have the potential to compete with J2EE someday, but they have a long ways to go. Zope + something roughly equivalent to EJB could perhaps work for smaller business apps.
I don't see how things have come out any better for Sun than if they had set Java loose.
.NET would have been the underdog, just like IIS. Now it's going to be an eventually open source "Java" trying to catch up to .NET's market momentum. Surely possible and a good goal, but it'll be harder now.
It would have led to broad and rapid adoption by the Open Source community. How successful has MS been at unseating Apache? If Java had been completely free and open from day one, it would likely have enjoyed massive market penetration. Once a really solid standard is in place, it's hard to shake it.
Quite frankly, the stupidest thing Sun did was force MS to give up Java. MS wanted to make Java ubiquitous by making it the standard platform for writing Windows apps.
It's not so much just what they did, but that they used monopolistic tactics to do it. MS was trying to completely take over the Java market, not just be an ethical player. In and of itself, there was nothing wrong with them writing their own implementation. (many others have done the same, including Open Source projects..) It was all the proprietary and undocumented extensions combined with their marketing power to sway the industry away from pure implementations without their extensions. This would have hurt us all, not just Sun.
How do we feel about MS's extensions to standards like Kerberos and HTML? Well, same thing.
First I'd like to express my opinion that Cringely is dead wrong. The war is far from over.
Linux needs to do something *groundbreaking* that Windows doesn't, that Microsoft can't suddenly copy, *and* that the public actually care about.
Absolutely correct.. I have three suggestions of things we can do that MS cannot duplicate:
1.) The Windows + Office duo is MS's cash cow. If you kill the cash cow, you quickly shut down the farm. The way to do this is OpenOffice, but OO.org needs a lot of help. Somehow, the Open Source community needs to figure out a way to fund OO.org and bring on a few dozen more full time coders to the project. Not forever.. just until it equals and begins to surpass MS Office. MS cannot give away Office on a large scale. It would kill their bottom line, they'd start reporting huge losses, and their stock price would plummet. And if the stock price plummets, those billions in the bank start getting used rapidly. For many people, the reason to run Windows is to run Office. Many large business workstations exist for the purpose of Office + web + internally developed apps. If you sweep the market for office suites, you get a lot of the market for Windows as well. They go hand in hand more than most people realize.
2.) Network booted workstations. We need to write tools that make it trivially easy to set up and administer a lab or office full of diskless, network-booted Linux workstations. The implications here for TCO are enormous for two reasons: One, you eliminate the most unreliable component of the workstation (cheap IDE disk). Two, the administrative costs fall dramatically because you only have to deal with the server. No more going around to 100 machines to install software updates, fix configurations, etc. Because of the licensing nature of proprietary software, this would be a difficult act to follow. Add Paladium and heavy-fisted DRM to the mix and it becomes near impossible! So even if MS started giving away Windows and Office, they'd still be losing the TCO war. And TCO is what most PHB's really care about..
3.) Business software. It's one thing to have a free OS and basic desktop software. It's quite another if businesses can rely entirely on free and easily customizable software for all their needs. Business software is typically not purchased from MS, although traditionally it requires a MS OS. So if that software is free and runs on a free OS instead, there's a much stronger incentive to ditch Windows than has ever before been presented. MS can't compete because this is software they don't themselves produce. Again, even if they give away Windows and Office, businesses still have to buy all the other software. The OSS community needs to fully embrace J2EE as the right tool for the job here. Then we need consultant-developers to take the software into the field, improve it, and interact with the user community. *hint hint* Anybody 'round these parts un/under-employed?
In general your points lack merit and an understanding of where the enterprise server market is today, but one point I will pick at specifically:
.NET is currently unable to really compete with it, from what I've seen. J2EE is massively popular and dominates its respective market the way Apache dominates web servers. In my opinion, J2EE needs to be fully embraced by the Open Source community for the purpose of developing free alternatives to high-end business software that hundreds of thousands of companies need (or would like to be able to afford!) Enterprise software is the final key to locking in the future of Open Source permanently. It's one thing to get a free OS and relational database. It's another when your whole business (or non-profit) can run on free, collaboratively developed, and easily customizable solutions. (Think of an "Apache" style project for all major business applications: accounting, ERP/CRM, communications, document management, etc.) For business purposes, the desktop is increasingly meaningless today -- it has largely become commoditized. (OS + office suite + web browser) Powerful, modular n-tier enterprise applications are the future.
"..and J2EE is too little and too late"
This could not be further from the truth. J2EE is a truly excellent solution for developing enterprise applications and
..copy-protection software this summer that will for the first time give portable digital music players access to rented tunes from all-you-can-eat subscription services
Right.. this will definitely fly.. because you know, it's not like the market is already saturated with 100's of millions of DRM-free home, car, and portable mp3 players.
but keep in mind that most PC users think MicroSoft is GOOD.
Is that why half the people I fix computers for have asked me at one point, "When is this Linux thing going to be easy enough for me to use? I hate this buggy Microsoft stuff that gets all these viruses and spyware and constantly needs fixing." At which point I usually tell them, "wait about another year and check back with me." I've been saying that for about 3 years now. But it's getting pretty darn close to the point where I can start making recommendations for some people to consider switching to one of the user-friendly Linux distributions. (People who have practical needs and don't care about the latest games and gadgets. They just want a computer that works without error.)
IBM have implemented their own clean-room JVM as of version 1.3. No connection back to Sun at all.
I've heard conflicting reports on this. Most people say IBM still has their hands tied. Do you have a specific, verifiable reference?
Here's my understanding of the scenario:
.NET is as to Java as Windows was to Unix.
.NET. Some would suggest Python, but I don't believe it has the potential for enterprise-class solutions. (yes, it's a great language otherwise.. just in a different league)
IBM should just open source their VM under a different license and not call it Java.
They can't or else they already would have done so. IBM's JVM is based on Sun's code and thus they have no rights to open source it. What IBM (and others) can/should do is financially support the Kaffe and GNU ClassPath projects, which are clean-room Java implementations written by people who have never been tainted by exposure to Sun's "community source" licensed JDK code.
As a sidenote, even Sun is legally unable to open source most of "their" Java code, so we really can't bash them for saying no. Third party contributors hold rights to much of what is contained in the official Java distribution. Sun really screwed up by not licensing Java under the GPL from day one, in my opinion. I believe they would have held just as much sway over the evolution of the language -- both by holding the trademark and by being the trusted leaders.
My prediction: Java will be the next Unix. It's a close historic parallel. Even more sickeningly ironic: perhaps Microsoft's
Or, of course, the other option is we (the open source community) invent something totally new that is neither Java nor
From what I've seen, Magnatune is by far the closest yet to doing things "right" when it comes to indie music. The total crap stuff is filtered and you're left with unknown but still professional musicians. (ie. no high-school garage band wannabe musicians) DRM-free + FLAC is the only reasonable choice for quality and compatibility. (yes, I'm an audiophile and quite picky.. If I pay for music, I want full CD quality, not 90%) And as for previewing albums, their streamed 128kbps MP3 is fine.
That being said, one big thing is missing: popularity. It's one thing to find music that I myself like, but there's still a missing link of how to get the best artists onto national tours where they can make a career-level income. So the question becomes this: will the indie music pay download sites enable sufficient popularity to make this possible? I have my doubts.
Radio + bootleg P2P is a powerful marketing machine for major label music, even if the RIAA has not yet acknowledged this. (which is the real reason you shouldn't share RIAA music!!) Always remember that the real money is made in concert. Whether somebody buys a CD or downloads via P2P, the only thing that matters is popularity. Popularity breeds concert ticket sales. So how can indie music compete with this?
It's a catch-22. On one hand, the musicians want income now -- so they try selling the music itself. On the other hand, they want to eventually become popular enough to go on tour and make a real living doing shows larger than bars and clubs. I'm not saying there's one answer for everyone, but I'm not aware of any bands yet who have fully embraced the marketing potential of unrestricted P2P and "Here's some free CD's.. Please make all your friends copies" Don't get me wrong, this would only work for top-notch bands in mainstream genres, but it seems the only way to compete with the major label marketing machine. For the less popular styles or for non-career musicians with no desire to tour, sites like Magnatune are probably the way to go. Hopefully an increase in independent (and online) radio popularity will help as well.
it's just simple market economics. i'd say driver availability pretty much follows relative market share.
Except that the "relative market share" here was not determined by market economics. Monopoly is still the driving force even if hardware manufacturers are only doing what economically "makes sense" to them. Incidentally, there is virtually no cost in releasing complete hardware documentation so that alternative drivers can be written. And yet most hardware companies still do not do this -- many still refuse, others release partial documentation that only allows basic functionality. This is the reason why many Linux/BSD drivers are so behind. Of course, in some cases, it would be ideal for the hardware manufacturers to help out with open driver development because so much is at stake for them. Video drivers come to mind because the volume of sales is significant. (ie. just about everybody buys either ATI or NVidia chips)
There are some software titles that just -need- to be ported to linux, do to lack of OSS alternatives.
Wrong. We don't "just need" any proprietary software to be ported to Linux. We do need to get behind the projects that are developing OSS alternatives and support them both community-style and financially. We also need to gather support of the business community, focusing on software that will save them money. ex.) "You spend $10,000/year on Macromedia tools? Support our project and you can drop that expense within 2 years."
I personally would gladly donate $100 to a professionally run project implementing SVG solutions so we can ditch Flash once and for all. I would donate more if it would give me a vote in future feature development. If we can buy Blender in a few month's time, we can surely pool enough resources to do this.
Exactly! Don't buy RIAA music. Download your sh** online
What you describe is only half-right in terms of solving the problem of a corrupt entertainment industry. The correct solution is this: Don't buy RIAA music but support independent / local / non-RIAA artists. That's right -- don't even share RIAA crap. Doing so only makes it more popular - and thus keeps people buying CD's and merchandise, watching MTV, and going to RIAA-artist concerts. And. incidentally, Hollywood is another good boycott target. Don't want DRM-laden HD-DVD's and HDTV components? Stop buying today's DVDs and going to every movie that hits the theaters! Cancel your ridiculous cable/satellite premium package! These people can do evil things only because YOU enable them with your dollars.
Look to software as an example. The answer to Microsoft's monopoly is not warez sites; it's Open Source. And it's working.
When alternatives exist to fight corruption, the legal one should be chosen first--not necessarily because the law is just, but because it's the easiest path. Unjust laws can be changed far more easily after monopolists have lost the reins.
I am with you on this one. Sun opens up Java and IBM does the same with their Lotus suite? A gesture is met with a gesture and we all win. It will never happen.
.NET is looming on the horizon, gaining acceptance fairly quickly. Sun needs all the help they can get.
IBM is not just saying, "Hey Sun, open up Java now!" as ESR did. They're saying, "We want to collaborate on an Open Source Java." There's a big difference. Read Sun's response to the ESR paper. Here's the relevant quote: "Sun would support an Open Source version of Java, but it need a lot of money and time to do so. You can't just flick a switch." In other words, Sun doesn't own all the code to Java so they can't just change the license. Other people have contributed code under the SCSL, which is not compatible with GPL. And Sun, being financially stretched as it is, does not have the resources to re-write all the Java code they don't control.
What IBM is saying is, "Yes, there is a significant amount of work to be done in bringing about a truely Open Source Java (ie. re-writing parts), but we're willing to help out and in fact bring most of the development labor" What Sun can do, if they're smart, is to begin releasing code that they control under a dual license just as with OpenOffice. And after IBM and others fill in the gaps, Sun can certify the code as meeting the spec.
Incidentally your example with Lotus suite makes little sense. Java is as much a piece of infrastructure software technology as Linux. Because of that, it needs to be truly open for maximum acceptance and innovation. Infrastructure is where Open Source has proved the strongest thus far. (OS's, libraries, compilers, server software, etc.) Although I do think IBM should open source Lotus someday too.
And what if Sun refuses to help with the Open Source Java effort? IBM should just go off on their own. They have the resources. They could bring together all the existing fledgling Open Source Java projects and unify them. This, of course, would not be beneficial to Sun and would likely lead to their demise, simply through diminished mindshare. It's really Open Source or bust.
There's no motivation to "Open Source" Java.
BS. The article gives several strong motivations, all of which I have personally run into as reasons NOT to use Java at this time. Java will not really take off until it is included with every Linux distro and can be fully embraced by the Open Source community. And it desperately needs more innovation, which OS community support would quickly provide. Sounds like you didn't RTFA. One thing the article didn't mention is that Sun's Java implementation is probably the worst in the industry.
It's more like a cow and some un-peeled potatoes.
But just think how many burgers you'll have once somebody teaches you how to use a meat grinder. (:
And everybody knows that real french fries are made with the potato skins on!
It will be interesting to see whether the Linux world can do a better job than Microsoft at providing security out of the box in order to protect these non-techie users from their own ignorance. Will Linux distros be able to stand up to countless virii targeted at Linux?
The design of Linux/BSD and the software culture of Open Source make for a quite unfertile environment for viruses and worms. Sure, it's possible that somebody could find a buffer overflow in a common userspace program like KMail and exploit it, but not everybody uses KMail.. not even a majority. (compare the majority IE/Outlook users). There's a bit more security paranoia as well. Feature bloat is kept to a minimum in areas that could potentially be exploitable with a discovered flaw. It also helps that, unlike with Windows, even clueless users of Linux/BSD don't use privledged accounts for daily use. Will we ever see workable malicious code? Possibly. But it won't be an internet-wide epidemic. There, however, is a case for further security-hardening Open Source desktops before they become more widespread.
By the way, "virii" is not a word. It's just viruses. (:
Let's face it. Linux isn't for just the uber-geek anymore. So logically, more systems are going to be hacked into when people with no security sense are managing systems. .. Don't blame the operating system. Blame everyone who thinks they're a competent sysadmin, but really aren't.
It's true, Linux is not just for geeks anymore. But because of that, we need pre-hardened distros (including ACLs, IDS, and stack protection) and automated security updates for systems run by idiots. The ultimate answer (educating people) is unfortunately not feasible. As much as possible, security needs to be idiot-friendly on every OS.
Intel is obligated to its shareholders to protect it's technology. Open source drivers could tip their cards to AMD or some newcomer could gain the upper hand. That is the REALITY of how the hardware business works.
BZzz.. wrong. You clearly have no understanding of how the hardware business works. #1.) It is common practice, when actually necessary, to reverse engineer a competitor's product to see how it works. Lack of source code for their drivers is not a hindrance. Hobbyists have in the past reverse engineered complex hardware in their basements in order to write open source drivers. Multi-million dollar engineering labs are more than capable. #2.) It takes too long to copy off somebody else's design. If you do, you'll be too late to market. #3.) Intel is obligated to its shareholders to make money, not "protect its technology." Sure, there are cases and ways in which it sometimes can/should, but this is not an example of one of them.
Do you reall think there are people who say "let's not do this open source project in Java because its not included with XYZLinux 9.3"
Yes. This sentiment runs fairly strong in the Open Source community. Many would like to use Java but do not because their software will then not be picked up by the major distros. This is a perfectly legitimate issue.
So are you referring to desktop Linux? That's such a small marketshare
It won't be a small 'marketshare' for much longer, and secondly, the use of Java for desktop applications is finally showing potential. Free Software on the desktop needs Java and Java needs Free Software on the desktop. Enterprise Java is going quite strong, but it's not enough to support Sun.
and again it's not like most Linux desktop users don't know where to get a JVM. So it's hard to see how there would be JVMs on Linux machines that currently do not have JVMs.
Wrong. Most of today's new desktop Linux users are clueless folk, not developers. They use whatever comes with their distro. If JVM is not included, they go without. Look at it this way: why isn't Java more popular on the Windows desktop? Because Microsoft doesn't include it by default! (of course this is for anti-competitive reasons, not licensing)
Sun should also consider this risk: If a third party Open Source project implements a compliant JVM and JDK, it will quickly become the dominant Java implementation on all Linux/BSD desktops and perhaps even beyond. (especially if it becomes faster than Sun's JVM with refinement) And it won't be called Java -- it'll just be compatible with the spec. So much for mindshare. This is the inevitable outcome in the next 5 years if Sun does not GPL as much as possible of their own implementation. Here's to hoping they do so we don't have to needlessly reinvent the wheel.
Personally, I don't think something like Java will gain any benefits from following the route ESR proposes. By setting the Java source code free will fragment it more than ever. And for an industry that needs to hold off M$ as much as possible, I think this would be a bad move.
The benefit to Sun of GPL'ing their Java implementation would be expansion of their market influence. Right now, there aren't very many open source Java apps (comparitively speaking). This would change rapidly if a complete JVM/JDK could be included legally with every Linux/BSD distribution. Complete adoption of Java by the Open Source community would mean a sharp rise in the popularity of the language and this would help Sun tremendously.
Keep in mind that if Sun GPL'ed their Java implementations, it would not mean a true loss of control. They would still own the Java and related trademarks. So even if somebody forked Sun's GPL code, it couldn't be called Java. And, in like manner, Sun would still control the specifications defining what "Java" is -- they would still have the right to certify what is and is not "Java". In reality, the situation would be no different than today, where 3rd parties are welcome to write their own Java implementations using the open specification.
So in the end, both ESR and Phipps are each right on certain things. But Sun has no advantage in keeping their JVM/JDK sources under a license more restrictive than GPL. The other question perhaps, is whether something legally prevents Sun from changing the license -- 3rd party code, etc.
I personally see Zope as the Free Software answer to JBoss, and I think that it is pretty darn good answer as well. As an application server Zope is a fairly amazing piece of work. However, it won't run your applications designed to run in a Java application server
.Net even though there is a small amount of overlapping functionality.
Both are of themselves just tools, but Zope is a completely different type of application server than JBoss. Zope is designed for writing web applications, not enterprise business applications. You would never write a CRM / ERP / Accounting application using Zope because it doesn't even have the beginnings of the requisite underlying functionality. Zope is good for portals, simple e-commerce apps, content management system, and the like. It is not a 'competitor' of J2EE/JBoss or
Somebody mod my post back down. I had received incorrect information regarding this issue. (Or was this a former Sun licensing policy that no longer exists?) My apologies.