ESR Responds to Sun's Claims of Being a Better Bazaar 310
UnixSphere writes "Sun has been quoted to have said, 'Sun's Java is developed more in the mode of the bazaar than Linux is,' which has prompted OSI President Eric Raymond to correct Sun's view of what open source really is."
Not sure about Bazaar, but it seems Bizarre (Score:3, Interesting)
JCP is anything but open (Score:5, Interesting)
Certainly a cathedral model.
OSS and Sun are on different sides (Score:2, Interesting)
Sun likes to cast these issues as "Sun+OSS vs. Microsoft" because it's good marketing, but that is an illusion and a lie. Sun helps OSS in some areas (which is nice), but with Sun Java, they have attempted an assault on open source and open standards. But Sun's assault is failing. The "cathedral" model under which Java is being developed is failing in the same way cathedral models have failed before: it's resulting in a bloated mess.
More debate! (Score:1, Interesting)
* windows vs. linux (vs. mac)
* vi vs. emacs
* creation vs. evolution
* republican vs. democrat (vs. independent)
I mean, that's all this kind of article can boil down to, so let's get the rest of today's arguments out of the way right now.
there are real issues at stake (Score:5, Interesting)
Sun is trying to market their products by taking advantage of the good will and trust that open source licenses have and misrepresenting their proprietary products as being associated with open source, and you blame "open source people" for it? You should be blaming Sun marketing and management. Their behavior has been reprehensible.
Open source people have better things to do than to worry about every single proprietary product out there. Get Schwartz and Sun to shut up about open source and cathedrals and bazaars and nobody will waste a second thought on Sun anymore. But as long as Sun keeps misleading people, open source advocates will respond because Sun's behavior is threatening the future of the open source movement.
Re:Not sure about Bazaar, but it seems Bizarre (Score:5, Interesting)
Here's how the concerned sides act to each other in a very simplified manner:
Open Source community about Microsoft: Shared Source isn't Open Source, but thanks for the instaler. Your closed source sucks because there are too few eyes.
Open Source community about Sun: It would be nice if you would decide where you really stand, but thanks for OpenOffice.org. Your closed source could be better with more eyes.
Sun about Microsoft: We would like to get some of the money you are getting from your monopoly-like marketshare, but you have shown that you can not be trusted.
Sun about Open Source: Your ideas are intriguing to me and I wish to subscribe to your newsletter.
Microsoft about Open Source: We like the BSD, we don't like copyleft.
Microsoft about Sun: Buzz off or we will crush you.
Re:I'm not sure how I feel about this (Score:5, Interesting)
Up to now, very few APIs have been proprietary. Sun has broken new ground by successfully asserting a high level of control over the Java APIs (not just their implementation).
If there is a fork, doesn't that present huge problems for the development community?
Languages like C, C++, Fortran, Perl, shell, and Python have all thrived in the absence of the level of control that Sun is trying to exercise. The reason is simple market economics: implementations that don't provide the features that users want disappear on their own.
Sun is trying to substitute their own interests for the wisdom and preferences of their end users. They are churning out one API after another, but users have no choice but to build on what Sun ships; even if there were alternative implementations, users would still be forced to accept whatever garbage Sun and the JCP dream up.
At least with C, you have the benefit of compiling. With Java, you are compiling to java bytecode, which is still interpretted, and still prone to problems between the forks.
Modern C programs have numerous shared library dependencies; Java's byte-code based system would, if anything, be more robust.
I guess you kind of experience this problem with shared libraries under *NIX, but at least you have the possibility for static compiling. You are stuck with the JRE for Java, no?
You are only stuck with the JRE for Java because Sun keeps you from having a choice. If Java were an open standard, there would be dozens of different implementations, and those implementations would work out amongst themselves what features were important core features and what features were vendor-specific extensions.
I don't believe that ESR (Score:2, Interesting)
Most Java developers have no intention of modifying or fixing the VM and are simply happy with the wonderful set of libraries available to them (Open source or otherwise).
As of 1.4, the quality of the Java VM has been ver good. JDK 1.5 rocks and the platform is alive and well. Thanks to Sun, IBM and mainly Apache.
Are things perfect? Not by any means. I just can not name one platform that I would substitue Java with to write my business applications.
Honesty (Score:3, Interesting)
I don't know the first thing about Schartz, so maybe he's just a slime ball or maybe he just didn't understand the underlying concepts of The Cathedral and the Bazaar, but this sort of behaviour seems to be considered fair ball play these days. And I think it is something that should be left behind on the playground. Heck, it wasn't that common on most the playgrounds of my childhood, outside of certain particular types of debates (where it was understood that different rules of conduct held sway).
Am I right? Is there more of this in the public sphere these days? Or is it just the same-old, same-old?
Re:JCP is anything but open (Score:5, Interesting)
Ok, let me get this straight...
Sun's model is cathedral like because you had to fill and fax a form?!
Re:I'm not sure how I feel about this (Score:3, Interesting)
I think the analogy is apt, but backwards. The Cobol/Fortran and C camps had mutual dislike. Cobol/Fortran represented entrenched, well-paid, proprietary interests. It was the analog of Microsoft today. C represented the slightly chaotic, open, non-proprietary alternative, like Linux today. And today, the dislike between Microsoft developers and OSS is also mutual.
Microsoft hatred is all about protecting the value of guild crafts and nothing about principle.
Yes, and that sums it up: people are tired of paying a premium for the Microsoft guild crafts, in particular since VB/VC++/.NET developers in general just aren't very skilled technically. That is why OSS has taken off. And OSS will beat Microsoft Windows and
On Java it was Sun who were being the evil proprietary monopolists. Their objective was to reduce every platform to the level of Solaris, leveling down, not up. Suns approach was "If you dare do anything that I can't I'll sue you."
Java could have been the future of computing but there is no way that any company, let alone a declining company like Sun can be trusted with the complete control they demand. The chances of Sun ending up in a SCO like position in five years time are significant.
I fully agree with those points. I think Sun is worse than Microsoft: Microsoft represents a particular approach forcefully, but at least they are honest about it (wrong, and doomed to failure, but honest). Sun, on the other hand, is just misleading people about what they are doing. And I also see the danger of an SCO-like meltdown. However, I think people are wising up to the threat and Java is becoming less and less popular for OSS.
Re:Free Forking? (Score:3, Interesting)
Sun went after Microsoft because they had a contractual agreement which stated they had to produce a product with certain attributes before they can call it "Java".
Sun has never prevented alternative Java implementation, there are many [java-virtual-machine.net].
As far as open-source there is Kaffe [kaffe.org], GNU Classpath [gnu.org], GCJ [gnu.org], Jikes [ibm.com] and others.
All those projects need help. And I am sure Sun is not the reason they are not getting it.
Put your money and time where your mouth is and support open-source Java
Re:Java (Score:5, Interesting)
This is interesting.
Java is prohibited from forking because Sun controls what is Java.
Linux is deterred from forking because it has the support of the community.
If Linux's management goes awry, then it will likely fork and spawn something with decent management. For example XFree86's X.org fork. This is good for the community. If Sun's management goes awry, the community must accept it. In other words, you are relying on the company to make the correct decisions.
I just thought this was an interesting perspective, and that I would share.
Re:OSS and Sun are on different sides (Score:3, Interesting)
Yes, I think it would. OK, maybe a few months behind. But if the whole community got behind KOffice, which is better integrated (Yes only with one desktop suite, but one is better than none, and for Linux on the Desktop to be a success you only need to get KDE accepted), less resource hungry, and ime (very little) nicer to work with, I think it would overtake OOo very quickly. It seems a better bazaar than OOo and does have efforts going on for a database frontend. While the spreadsheet is somewhat lacking, I can see the gnumeric backend having been integrated when enough people were working on KOffice to make a difference.
OOo has two advantages over KOffice, name recognition and ports to non-nix systems. Neither of those are critical to Linux on the Desktop. So in summary, yes, I think if OOo didn't exist KOffice would have got a large number of the OOo coders and actually be more useable now than OOo currently is.
Re:This exactly matches democracy vs free markets (Score:3, Interesting)
Re:Java (Score:3, Interesting)
It is really disappointing to me that the GPL-advocates insist that Java isn't free and/or open-source.
That's because it isn't free or open source. It would be kinda cool if it was, but it isn't.
1) People assume that Apache is open-source
That's because it is. Not all the stuff is GPL compatable, but it is Open source.
2) Apache has a single point of control
The Apache project has the most influence on Apache projects, yes. But they don't control their projects the way Sun controls Java.
3) The JCP has an elected board that is in control
4) Apache has as much control as Sun on that board
Lots of Sun people work on Apache projects, but so?
5) Patches that I have submitted to the Ant project were never even discussed, let alone integrated
Just because a project is open source, it doesn't mean that they are obligated to accept your patches. You can either (a) gently continue to pester them and get other people to do the same, or (b) start your own build program. Call it 'antz' or something more creative. You are allowed to do that under the license, I just checked.
There is more I could reply to in your post, but most of that falls under "So what?".
This is a gross oversimplification, but if you can take the code, and make and release your own version, then it is open source. If you can't, as with Java, then it isn't. I don't know why you don't seem to understand that. Or maybe you're a troll, and I'm a fool for responding to you.
As a side note, I can tell you that if Java were GPL'd, it would not have been used by any company I have worked for. They ALL had a strict policy against using GPL'd software.
Well, if your companies want to hurt themselves, and not avail themselves of the best technologies possible by limiting their choices, that's fine. It's a free country. Kinda. Sorta.
Re:Java (Score:3, Interesting)
Right.
But, in the eyes of Sun, new implementations should be worse than forks based on the same original source, right?
Wasn't that obnoxious Microsoft "fork" a new implementation, as well?
Re:Java (Score:3, Interesting)
2) Apache has a single point of control
3) The JCP has an elected board that is in control
4) Apache has as much control as Sun on that board
Did you read the FA? Can you fork Apache, thus taking the source and start your own version? If yes, then it's open source, if no, then it isn't.
5) Patches that I have submitted to the Ant project were never even discussed, let alone integrated
Same here. Can you say to the Ant project to shove it and start YourAnt, using the current development source tree?
6) Pathces that I have submitted to various Java projects have been implemented
Accepting or rejecting patches does not make stuff more or less open source. Again, if you can fork the source code of these project, thus taking the current dev tree of the project and start your own branch, it's open source. Otherwise it isn't.
7) The only advantage Sun has in the JCP is that it has a permanent seat on the board
8) You __insert_name_here__ can have as much say as Sun if you run for the position
9) The JCP *requires* open-source members on the board
10) This last election, JBoss nominated themselves and won -- http://www.jcp.org/en/press/pmo/04openletter
You might have guessed. Can you take Sun's implementation of Java and fork it? No? Then Sun's Java implementation is not open source and you are at the mercy of their continued support of the product. This has nothing at all to do with JCP, as JCP as far as I know does not develop any software. It's a committee designing a language, and as such are completely outside of the discussion of open and closed source. Java can be open source, it's simply that Sun's Java is not. You can hold Sun in the highest regard for their support of the industry, guarding the Java language against the beast of Redmond, and all that, but that does not make their source open. When and only when there's a certified implementation of Java that is open source, can Java be called an open source language. Until then it's a proprietary language with decent license conditions.