Apache sends open letter to Sun over license terms for TCK

Home

News: Apache sends open letter to Sun over license terms for TCK

  1. Geir Magnusson, Jr., has posted an open letter to Sun documenting Apache's inability to "acquire an acceptable license for the Java SE 5 technology compatibility kit." The license for the TCK includes a limit regarding field of use, which the open letter says is contrary to the spirit of open source and yields a significant commercial advantage to Sun.
    Besides holding back the Harmony project - a community-led open source project of the ASF since May of 2005 - this failure to comply with your contractual obligations poses serious risk to the credibility of the JCP as an open standards organization, and the reputation of Java itself as an open technology. We believe that this also threatens the general cooperative nature of the commercial Java ecosystem, puts at risk the long-standing positive relationship between Sun and the ASF, and probably between Sun and the broader open source community - all of which is key to the continued growth of Java. Beyond the obligations of the JSPA, these limitations are also contrary to Sun's public promise that any Sun-led specification would be fully implementable and distributable as open source/free software. It shouldn't have to be mentioned that "fully implementable" includes passing the JCK, as required by the specification license. To this end, limitations on field of use for our users is contrary to the basic principles of open source licensing, and therefore these limitations would prevent distribution under any open source license, including our own. Our objections to the offered license are clear and valid. The situation we are facing is grossly in conflict with the basic IP philosophy of the JCP, the concept of Java as an open standards-based ecosystem, Sun's public promises to the free and open source communities, and Sun's contractual obligations as a specification lead under the JSPA. The JCP was clearly designed to prevent any single actor from being able to exhibit this sort of market control. Additionally, it is contrary to both the spirit and letter of open source, the respect of which is a key element in Sun's stated business strategy. Through Apache Harmony, the ASF is implementing Java SE in good faith, with the understanding that Sun as the specification lead will reciprocate. Our intention has always been to produce a certified compatible implementation of Java SE distributed under the Apache License. To do so, we need the JCK.
    Sam Ruby posted "JCK Access," offering his thoughts:
    I sincerely hope that Jonathan quickly intervenes as he is in a unique position to assess the trade-off between the short term benefits in the credit column against the intangible costs in the debit column of (1) actively destroying the community that Sun has taken so much time and effort to foster, (2) mortgaging the future of Java, and (3) undermining Sun’s own open standards efforts. Specifically:
    • Harmony is exactly the sort of responsible, community-led evolution that the JCP rules were changed to facilitate. More background on the events that lead up to those particular changes in the JSPA can be found here.
    • Down the road, I can see no way that the Executive Committee for Java SE/Java EE would approve a JSR for Java™ SE 7 Release Contents as long as the possibility of the spec lead tacking on such noxious FOU restrictions still exists.
    • The JSR 270’s spec lead’s proposed FOU restrictions conflict with Sun’s own Open Standard Definition.
    Finally, it is my understanding that a number of Sun employees have attempted to help, but were blocked by people higher in the pecking order. Their efforts are most appreciated. Keep up the good fight!
    It'd be interesting to see why the fields of use restriction is still in the compatibility kit license, or what the justification is. The open letter gives a 30-day deadline, without actually clarifying what the potential response is if the deadline passes; this makes it less than an ultimatum but it also communicates displeasure. Sun?

    Threaded Messages (28)

  2. sun afraid from IBM?!!!![ Go to top ]

    the whole harmony project was inspired and led by ibm which was the first to ask publicly to opensource java and when sun refused it helped apache for the implementation by donating the initial codebase of the classlibrary and the testing JVM for harmony ,and sun think that as ibm was successfull in making Geronimo(websphere community edition) a much more popular and deployed appserver than glassfish, it will replay the game again in supporting Harmony versus openJDK,leading to losing the control of the java system to IBM isn't that true sun????
  3. re: geronimo[ Go to top ]

    Unless you are going to pull a Stahl and post dice.com numbers for developer opps., i would suggest you demonstrate how geronimo has been a success vis-a-vis glassfish... The reality is that IBM is more committed to SCA than to Java at this point, they have not come out with JEE5 support in Geronimo, and may never with WebSphere proper... I doubt there is much concern for IBM at Sun over Java, unless you consider a development environment to be the cornerstone of business with respect to JEE, which i don't agree with... As for a JDK platform, how does it really matter? I remember the day that Sun announced open source Java, and there was some ceremony behind it, but it never garnered the break-thru that some at Sun point to... Whether there is a successful openJDK v. Harmony does not seem to matter much, is it really that likely that there will be enough support to make Harmony the default?...i don't really see it... apache is grasping, ASAIC...
  4. Re: re: geronimo[ Go to top ]


    The reality is that IBM is more committed to SCA than to Java at this point, they have not come out with JEE5 support in Geronimo, and may never with WebSphere proper...

    I doubt there is much concern for IBM at Sun over Java, unless you consider a development environment to be the cornerstone of business with respect to JEE, which i don't agree with...
    I think it's fair to say that IBM's strategic direction is SCA - but it's hard to see how Java (if not JEE) can not be integral to their plans. I think that too much is read into the fact that SCA was not part of the JCP. It's because SCA is a SOA specification - so it needs to be language neutral. The evidence for this is that IBM and the other SDO backers took the trouble to put the Java SDO specification back into the JCP (and the C++ version into OASIS, like SCA). Sun has now joined the SCA/SDO effort..... so just about the only major vendor missing is Microsoft (which has its own competing WCF).
  5. Re: re: re: geronimo[ Go to top ]

    PJ, I understand that Sun is part of the SCA, but as I have outlined before it seems to be a half-hearted effort, considering it runs in direct conflict with some of their most important assets, namely JEE. I don't see SCA circumnavigating around JBI or JAX for Glassfish, and Glassfish will be a major app server platform, as long as Sun is alive... As for IBM, they most definitely have very little concern about the ramp-up of JEE5, for they simply do not have the type of customer that clamors for the best spec. available, they just look for the best recommendation from Global Services, therefore, as SCA becomes a serious challenger to Java on SOA, they will invest Global Services in "pure" web services, irrespective of JEE5... This much is clear to me: Geronimo will be the first platform that will incorporate SCA as primary development environment for web services...it will not be JAX, it will be Eclipse with OASIS standards...because in the end, IBM has to do exactly what I said before, both of them: re-write everything according to Geronimo, and dominate with SCA, thats it... that appears to mean even the re-emergence of C++, sorry I'll bet my money on Java and Sun on this one...
  6. Re: sun afraid from IBM?!!!![ Go to top ]

    ibm was successfull in making Geronimo(websphere community edition) a much more popular and deployed appserver than glassfish
    I take it you have never deployed/used (websphere community edition). I'm not defending glassfish, but considering websphere community edition a successful product is going too far.
  7. Come on now![ Go to top ]

    We should definitely wait for Sun's response. As for IBM, do you think that any commercial entity would be so philanthropic? I would say from my own limited vantage point that IBM probably gains more revenue in terms of software than Sun does, meaning they are aggressive in their hardware/software revenues vs. being more like a hardware company that sells software, I could be wrong. If IBM were truly, meaning sincerely, vested in Geronimo, than for me at least, they would have to put their money where their mouth is and make Geronimo/Websphere community edition the base for _ALL_ of the websphere stack, instead of the gateway drug for an upsell (I could be wrong here, but last I checked it wasn't someone please correct me if its inaccurate).
  8. Re: come on now[ Go to top ]

    I am 99% sure that WebSphere is in no way related to Geronimo except for community Edition, and the entire WebSphere code base is legacy code that is a hold-over from as far back as '99...is this the same as WebLogic, as well?...i'll lay off BEA, though since they're not in this debate, as yet... The WebSphere stack is a SOA environment waiting to happen, with integration points that span the spectrum of literally everything, second to none. I believe that their only way out of this mess is two options: 1) re-write everything according to Geronimo 2) adopt SCA to justify everything Anything to do with Apache and Harmony is a distraction that continues the myth that they are more committed to Java than Sun, this is no longer true, with all the proprietary extensions, the lack of JEE5 roadmap, and elimination of JBI and JCP as the main extensions of the ESB effort... They are in the same boat as Oracle and BEA, recognizing that Sun has an app server program to consider, and not wanting to go toe-toe with them, they obscure the truth, it is a shame, it is pure-Microsoft...
  9. Re: come on now[ Go to top ]

    Actually, if I were responsible for strategy at IBM I would be petrified at Glassfish. It has the potential to be incredibly disruptive to their revenue stream, especially considering that retail pricing is around 12-13% for support (Sun App Server/Glassfish vs. Websphere/NON-community edition). I would ask others in the community to just reflect on Douglas' post, and past empirical behavior around what IBM has done (anyone remember Soap4J-->Apache Soap-->Axis-->IBM pullout?). I think the tendency is more disingenuous than some may think, meaning, it all ends up in Websphere, the part you have to pay (a lot) for, and some times, not even the versions that are GA yet. If I could swap out our Websphere licenses I could hire more developers to actually ease the bottleneck in our delivery stream. Instead, decisions made at a much higher level has our hands tied into ELA's that chew up the budget. I'm sure I'm not the only one who works in a corporate environment that has experienced this. I'm also not talking from a vantage point of no-touch with this stuff (I've implemented Websphere from version 3.5, 4, 5.1, 6.1 now), and I have yet to meet/talk with anyone who is passionate about what they are doing who can say that they "LOVE" Websphere. If you do, let's start a campaign and petition list, and if you don't, voice it here, and maybe it will be loud enough for decision makers to listen beyond the "you won't get fired for buying IBM" mentality. I'm not trying to flame the Websphere guys/gals, lots of great people, however, once it goes into procurement, governance, budget, SALES PEOPLE, it makes my life more stressful, especially as there are better, cheaper, alternatives in my very opinionated opinion ;-)
  10. Re: come on now[ Go to top ]

    I have yet to meet/talk with anyone who is passionate about what they are doing who can say that they "LOVE" Websphere.
    +1. My company builds JSF UI development tools for the Dreamweaver platform. I was shocked to learn that IBM explicitly prohibits sharing WSAD JAR files outside of their tools: Legality of Using IBM runtime or build jars outside of WebSphere Studio IBM has added a proprietary compile-time dependency (jsf-ibm.jar) that can only be found in their IDE. They have raised an artificial barrier between WSAD and other development tools to try to prevent round-trip engineering. Talk about vendor lock-in! Ian Hlavats JSFToolbox - JavaServer Faces for Dreamweaver
  11. Funny that this was posted a day after I found this on the java.net blogs ... http://weblogs.java.net/blog/alexeyp/archive/2007/04/compatibility_t.html ... S.
  12. Knew this was coming[ Go to top ]

    Limiting the field of use is indeed against the spirit of Free Software. I hope that Sun does make a change. Honestly, does anyone have legitimate fears of Sun somehow losing control of Java? I sure don't. I don't think Sun has much to fear by making the change. -John Mark
  13. Re: Knew this was coming[ Go to top ]

    Limiting the field of use is indeed against the spirit of Free Software. I hope that Sun does make a change.

    Honestly, does anyone have legitimate fears of Sun somehow losing control of Java? I sure don't. I don't think Sun has much to fear by making the change.

    -John Mark
    Is it? From a pragmatic and philosophical point of view I don't see how this is any different from the GPL restricting what kind of software you can link to GPL software. Personally I agree, it is against the spirit of free software, and so is the GPL. But there are an awful lot of people who disagree with me, including the people who coined the term.
  14. Re: Knew this was coming[ Go to top ]

    Limiting the field of use is indeed against the spirit of Free Software. I hope that Sun does make a change.

    Honestly, does anyone have legitimate fears of Sun somehow losing control of Java? I sure don't. I don't think Sun has much to fear by making the change.

    -John Mark
    Hyperic


    Is it? From a pragmatic and philosophical point of view I don't see how this is any different from the GPL restricting what kind of software you can link to GPL software.
    I see what you're saying, but I see a big difference. The principle of the GPL is to mandate that derivative works are just as open as the original work, although there are certainly quirks that result from this philosophy. You noted one of them, the inability to link in all types of code. In this case, however, it would appear that "limiting use case" is designed to limit competition and prevent forking. You seem to be saying the 2 cases are equivalent and I disagree, because the intent of each is quite different. -John Mark Hyperic Community Manager
  15. Two sides to every story[ Go to top ]

    I think people should wait for Sun's response before passing judgment on this. The letter + the FAQ make Apache's claim seem very reasonable. Perhaps it is, or perhaps their distorting the situation. Anyway, this reaks of IBM. The community may gain significant IP if Apache wins, because I bet IBM will start pouring resources into Harmony. But I also think IBM will use it as a ploy to usurp Sun's influence over Java and kill Sun's Java revenues. I think this bodes ill for Java.
  16. 期待sun的回复[ Go to top ]

    对于这个事情,我只想看到sun的回复,让harmony好好发展
  17. Good bye Harmony...[ Go to top ]

    I hope Harmony join to SUN open source distribution (or just quit).. two different JVM distributions will take Java in the same trap as Linux.. were thousand of people keep redoing the same tasks in different ways... for me, good bye Harmony .. forever..... it was a good dream..
  18. Apache Geronimo is very different from IBM Websphere Application Server. It is a product bought from GlueCode by ibm. IBM Webphere CE is a free(no fee) as based on Apache Geronimo, but it is no ralated to the orginal Websphere product.For some makert perposal ,IBM named it "Webphere CE" to confuse customer to determine their purchase. Everybody know that IBM control the project Apache Harmony. When Sun opensourced Java under GPL, everybody enjoy it but IBM. Obviousely IBM want to control java througth Apache Harmony. I dislike it.I beg IBM give up Apache Harmony and focus on Sun OpenJDK project.
  19. Hum, Apache being able to certify that Harmony implementation of Java EE 5 is valid is, AFAIK, the matter of this open letter. Indeed, it's what the TCK is made for : assert some Java version compatibility. Apparently, the ASF thinks that the licence of the TCK doesn't allow them to certify Harmony wih the Apache license. However, in their FAQ, I didn't see how exactly it was "forbidding" it, just some general comment about the use made of the TCK approved material but no real quote of the part of the licence involved. At the end of the day, it could just be a stupid leftover of older license which can be pretty easily removed, or something more serious, I don't really. What is sure is that it won't give the ASF control of Sun implementation of the Java specifications. I'm not even sure the ASF wishes so ! Cheers, ZedroS
  20. There's been quite a bit of IBM talk in this thread. I think it's probably worth pointing out that Geir (the person who sent the open letter to Sun) has a job that doesn't pay him to work on Harmony, and previously worked at Intel, where his paid work on Harmony was incidental to his responsibilities. He hasn't worked at IBM for a little while now, and when he did, it was because of the Gluecode acquisition, which has nothing to do with Java SE or Harmony. There is no conspiracy here.
    Obviousely IBM want to control java througth Apache Harmony.
    I don't see how this news item, or really anything else to do with Harmony, can lead to this end goal. Both Apache and Geir have been pretty clear about their support for Java-the-brand from a compliance standpoint. In fact, this very news item is about Apache trying to convince Sun to allow Harmony access to the compliance suite under terms compatible with the JSPA (and incidentally, with the Harmony implementation). Finally, even if there is some conspiracy, I don't understand how IBM could hope to wrest control from Sun through an Apache Licensed source base without control of the brand, TCK and RI. -Patrick -- Patrick Linskey http://bea.com
  21. control of Java[ Go to top ]

    Patrick, I don't want to be associated with your so-called 'conspiracy' just because I have an opinion on IBM's position with respect to Java...I may have been over eager to dismiss this thread as a lame Apache attempt to help IBM, but its tough to see it otherwise, when openJDK exists...what would you expect from Sun today but a punt on whether they will address Geir's concerns...I'm not impressed by that blog entry from Sun, but when you call them out publicly, and claim that all avenues have been exhausted outside of the most inflammatory route, then you are going to get a non-comment... My question is why would IBM want to be the default compatibility expert, other than to sway things in Geronimo's direction, so that if Harmony was the SE environment that somehow became the default for the industry, would the community accept anything that smacked of preferential treatment...let alone do something like include SCA in the code base and regulate JBI/JAX to EE, only... I don't think anyone cries conspiracy except those who are only interested in listening to positions of commonly-held wisdom, because once they hear an alternative viewpoint, they consider it to be a plea for someone to look in to the so-said conspiracy...no one is crying conspiracy, just pointing out the obvious: IBM could benefit from Harmony, and Sun knows it...
  22. Re: control of Java[ Go to top ]

    I've been trying to stay out of this thread, but this one just needs an answer... So here's one reason why people are using the word 'conspiracy' : Douglas Dooley :
    My question is why would IBM want to be the default compatibility expert, other than to sway things in Geronimo's direction, so that if Harmony was the SE environment that somehow became the default for the industry, would the community accept anything that smacked of preferential treatment...let alone do something like include SCA in the code base and regulate JBI/JAX to EE, only...
    You think that Apache wants to get the TCK for Harmony so IBM, who already has their own TCK, can.... include SCA in the codebase to sway it to Geronimo and regulate JBI (which they hate) to EE (which the make lots of money on) only? Really?
    I don't think anyone cries conspiracy except those who are only interested in listening to positions of commonly-held wisdom
    I'll be the first to admit your theory about SCA, JBI, Harmony and the TCK is definitely classified as not being "commonly-held wisdom".
    IBM could benefit from Harmony, and Sun knows it...
    This is a true statement. Of course, lots of people can benefit from Harmony, but don't let that fact ruin a good conspiracy theory... ;) geir
  23. Re: control of Java[ Go to top ]

    Jnr, Part of your stance is down to how unsuccessful Gerongo has been when compares to Sun's implementation. Spend your time working on Gerongo and documenting it. Have fun!
  24. Re: What's with the IBM bashing?[ Go to top ]

    Venting ;-).
  25. A question to Geir?[ Go to top ]

    Hi Geir, What is the purpose of Harmony to implement a JDK 1.5 implementation? Sun already release JDK 1.5 for 11 updates, and Harmony currently just 95% complete but not compatible with JDK 1.5. It may take another 1 year to complete it implementation, who knowns how long it will take to be mature. (stability, performance, compatible ...). My question is, since there is a much mature JDK 1.5 implementation there (for several years), and supports most main stream platform, and GPLed (more open than APACHE license), why are you try to reinvent the wheel? Your wheel is more round?
  26. A question to Geir?[ Go to top ]

    Hi Geir, What is the purpose of Harmony to implement a JDK 1.5 implementation? Sun already release JDK 1.5 for 11 updates, and Harmony currently just 95% complete but not compatible with JDK 1.5. It may take another 1 year to complete it implementation, who knowns how long it will take to be mature. (stability, performance, compatible ...). My question is, since there is a much mature JDK 1.5 implementation there (for several years), and supports most main stream platform, and GPLed (more open than APACHE license), why are you try to reinvent the wheel? Your wheel is more round?
  27. Re: A question to Geir?[ Go to top ]

    why are you try to reinvent the wheel?
    Actually, the 'why' is unimportant in this dispute. Java SE 5 is a (supposedly) open standard defined by the Java Community Process (JCP). As such, Apache Harmony (or any other person/group) has every right to implement the specification, regardless of 'why' they want to. The dispute is centered over Sun's license terms for the testing kit needed to prove that the specification has been met. Sun has added terms which prevent the ASF from taking that test kit - terms which are against the JCP (JSPA) contractual obligations. Stephen Colebourne, ASF member, speaking personally
  28. Re: A question to Geir?[ Go to top ]

    why JoNAS and Geronimo and Glassfish when JBoss is there? why openESB when serviceMix is there? why cxf when axis is there? why activeMQ where openJms is there? complete an endless list.... Harmony is there because some people needs it AND because sun promised that that will be possible and it must keep itself some cridibility? harmony IS the project which pushed sun to opensource its implementation and it is there to ensure that the platform will remain open and available everywhere(that was the java promise,is't it) another reason that obviosly compition is always good and will definetly benefit you yourself if u donot like it ,donot use it but the main players are looking seriously to it and backing it(ibm and intel are major contributor and bea is an early adaptor as it poffers a JDK with the harmony classlib) for all who wish that harmony die: THIS PROJECT IS FOR YOU please support it or at least donot hurt people who wanna help u
  29. Re: A question to Geir?[ Go to top ]

    Wayne Zhang :
    Hi Geir,
    What is the purpose of Harmony to implement a JDK 1.5 implementation? Sun already release JDK 1.5 for 11 updates, and Harmony currently just 95% complete but not compatible with JDK 1.5. It may take another 1 year to complete it implementation, who knowns how long it will take to be mature. (stability, performance, compatible ...). My question is, since there is a much mature JDK 1.5 implementation there (for several years), and supports most main stream platform, and GPLed (more open than APACHE license), why are you try to reinvent the wheel? Your wheel is more round?
    I'm not really sure how to answer this. There are a bunch of things here. Having more than one implementation of specifications is *good*. It gives users choice, and tends to drive improvement. Look at Java SE performance - I'd argue that having other implementations of Java SE from IBM and BEA has led to great performance benefits to users - IBM, BEA and Sun are constantly competing for performance bragging rights, and that means we users get faster Java. It also gives users power in our relationship with our vendors - we can always choose to change. Also, having multiple *independent* implementations of a specification is wonderful for users in terms of price and innovation. Right now, both IBM and BEA (and many, many others) license parts of their Java SE implementations from Sun (under commercial terms), which is good and bad. Good, because there's a common codebase to work from, but bad, because there's a chokepoint for implementation innovation. For example, Harmony has a cleanly modularized class library, that has all sorts of already-proven benefits. Imagine if one of Sun's classlibrary licensees wanted to do that. They could either try to maintain it themselves, re-doing all the modularization work every time they get a new code drop from Sun, or they could hope Sun would take those changes and apply them to the codebase. (History has shown that Sun is very reluctant to do things like that from licensees). So in short, just because Sun's java SE 5 exists, having Harmony do an independent implementation is still good (we'll do Java SE 6 as well, of course). Even if you don't see any value in it, it's ok - it's work that *we're* doing, and it doesn't harm or impact anyone else. The fact that OpenJDK will exist under a the GPLv2+CPException has no bearing. Just like JBoss, Geronimo and JOnAS didn't throw in the towel when Sun finally decided to open source the Java EE RI in the Glassfish project, Harmony isn't going to quit simply because Sun finally decided to open the Java SE RI in OpenJDK. Finally.... I don't want to turn this into a license discussion because Apache's argument with Sun is about Sun's contractual obligations as a spec lead, not open source licensing, but I just couldn't let the GPL/AL comment go... do you understand the difference between the GPL and the Apache License? Claiming that the GPL is "more open" is probably the most compact, most Orwellian bit of license confusion I've ever seen. You do realize that you give up significant rights to how you can license your own code that you created when you accept the GPL and combine code under the GPL with your code? That's "more open"? geir