Discussions

News: Sun fires back over Open Source Java accusations

  1. Sun fires back over Open Source Java accusations (28 messages)

    Sun has offered a frank response to the open letter from Eric S, Raymond, President, Open Source Initiative, in which he called on Sun to make its Java platform Open Source and described the company's Open Source strategy as 'spotty' and 'confused'.

    'I'd say this is 100 per cent rant,' Sun's Chief Technology Evangelist, Simon Phipps told us. 'His simplistic accusations don't hold water... If this is the way that Open Source treats its friends, I'd hate to see how it treats its enemies.'

    Read Sun fires back over Open Source Java accusations

    From the blogging world

    Why Sun Hates OpenSource According SimonPhipps

    Eric Raymond's Letter and Modern Computing Mythology

    Simon Phipps on ESR response

    Threaded Messages (28)

  2. Am I missing something, or there are already some Open Source implementations of the JDK (Kaffe, for example)? What is the problem regarding the creation of an Open Source implementation, besides the huge effort? Are there any kind of legal impediments?

    OTOH, if they are actually talking about open-sourcing the specification of java, putting it on the hands of an "independent" entity and not a single private company, my view is that the JCP, even not being a the perfect process, still allows almost the same kind of openess seen on OS projects. It is not the same, but at least people have some kind of saying, and the process have improved lately, and may become more and more open, thus making this open-source movement even less critical.

    I sense some confusion is going on regarding the difference between the "implementation" and the "specification" of Java, at least.

    But then, I may be completely wrong, what is highly probable.

    Regards,
    Henrique Steckelberg
  3. Am I missing something, or there are already some Open Source implementations

    > of the JDK (Kaffe, for example)?

      To find out the most popular free Java-like VMs I've kicked off a Java Republic online poll. The line-up includes:

      * Kaffe
      * GNU Compiler for Java (gjc)
      * Mono (+IKVM.NET)
      * Sable VM
      * Kissme VM
      * Joeq VM
      * SuperWaba
      * IBM Jikes Research VM (RVM)
      * Intel Open Runtime Platform (ORP)
      * Other (Please Comment)

      I invite you to cast your vote online @ http://viva.sourceforge.net/republic/2004/02/poll_is_kaffe_hot_or_not_what_is_your_free_java_vm_of_choice.html

      - Gerald
  4. What is the problem regarding the creation of an Open Source implementation,

    > besides the huge effort? Are there any kind of legal impediments?

    Allow me to quote from the Debian GNU/Linux Java FAQ online @
    http://www.debian.org/doc/manuals/debian-java-faq

    Q: What are the problems with Suns' new license? A: Sun has moved to a new license the Sun Community License, like the GPL it is a viral license, but making all it touches subject to Sun licensing fee. The SCSL even goes so far as to define any implementation of a Sun specification as a "Modified Work". Basically, this means that if you implement any part of the new 1.2 API or Jini API, even from scratch, Sun will "own" your implementation and you will have to pay them for the right to use it.
  5. We did not feel that Sun hates us at all!

    In the process to get our Open Source project (eXo platform) a JSR certifcation (portlet API), we were in contacts with Sun Microsystems. And everything was fine from the quickness of the answers to the quality of the support.

    I guess this is the same with the Object Web consortium for the J2EE 1.4 certification.

    And there is also JBoss now...

    It makes more and more Open Source projects that will get a Sun Microsystems certification...

    Don't expect too much, too fast...

    Benjamin Mestrallet
  6. I dont understand what OSI is exactly looking for.

    So far, the JCP process and java itself is largely free comparing with other licenses.
    For example, Sun Java implementation and runtime is free to use in production code , and BEA offers JRockit.
    Only thing is, you dont get the source code. But how many people do really need source code?

    I even heard Red Hat is planning to ship JDK as part of their operating system. Is not this sufficient?
    sombody enlighten me!!
  7. why the source code is important[ Go to top ]

    Only thing is, you dont get the source code. But how many people do really need source code?


    The source code and an open license (like GNU or BSD-style) is important, because they give a certain guarantee that the software will remain open. If BEA for example has developed JRockit, they have all the IPR rights for it. They may distribute it for free at the moment, but nothing prevents them from changing it to a commercial product at any time.

    Licensing the software under GNU license and distributing source code does not prevent them from making the move described above. However, it is very likely that some other organization picks up from where they left and continues maintaining an open source version of the product.
  8. Only thing is, you dont get the source code. But how many people do really need source code?

    >
    > The source code and an open license (like GNU or BSD-style) is important, because they give a certain guarantee that the software will remain open. If BEA for example has developed JRockit, they have all the IPR rights for it. They may distribute it for free at the moment, but nothing prevents them from changing it to a commercial product at any time.
    >
    > Licensing the software under GNU license and distributing source code does not prevent them from making the move described above. However, it is very likely that some other organization picks up from where they left and continues maintaining an open source version of the product.

    Aren't we confounding up the "implementation" with the "specification" again? If Sun decided to throw away it's JDK, it wouldn't be such a big impact, since there are IBM's, BEA's and other Java implementations out there. Open Sourcing the JDK woudn't mean a big thing, as we already have many options besides Sun's own JDK in the market right now.

    The real problem is Open Sourcing the "specification", to transfer JCP control from a private company to an independent body.

    Regards,
    Henrique Steckelberg
  9. hackers are the wrong market[ Go to top ]

    Who cares what the hackers think? What I care about is what the consumer thinks, that is where the market is. Java has enough technical mindshare that we know it is not going away. But what it doesn't have is penetration into the consumer market. Thats what it needs now - not a bunch of hacker tools that will let some geek in a backroom do a DB2 backup faster.

    Considering how unfriendly many OS projects are, hackers are not going to be the creators of those mass-market programs that the average, non-technical, f*k-u-and-your-RTFM-attitude consumer wants and pays companies/developers for.
  10. What?[ Go to top ]

    Java developers have been against vendor lock-in all along. Now you are saying that it doesn't matter that a single company controls an entire language?

    Every other major language C, C++, C# (really), Perl, Python, Ruby, etc. have implementations of their compilers/interpreters in the public domain as open source. The response from Sun that this doesn't apply to Java because it's not a 'scripting language' is nonsense.

    Do you care about standards? Do you care about open source implementations? If you do then you should be supporting the push of the open source community to encourage Sun to get with the program. C# is already a standard under both ISO and ECMA with an open source implementation.
  11. What?[ Go to top ]

    Do you care about standards? Do you care about open source implementations? If you do then you should be supporting the push of the open source community to encourage Sun to get with the program. C# is already a standard under both ISO and ECMA with an open source implementation.


    I do care about standards, and I think that Java should be submitted either to ECMA or ISO, but how much of Microsoft's push to 'standardise' C# was really to do with legitimising the language rather than marketing, one-up-manship, and ecolumn inches? Also just remember that the language spec being standardised means very little to a developer.
     
    1) The language changes are infrequent so changes, are in general manageable regardless of who is responsible for them.
    2) The meat is in the class libraries (I can't see MS opening up much of the System namespace anytime soon)
    3) Even for languages such as C, COBOL, and SQL, all with several standard revisions under their belt, we have custom pieces added on (look at MFC Vs. Borland's VCL, or OWL for that matter) by the vendors.
    4) Do Novell/Ximian and the Mono developers actively have input to the evolution of C#?

    If it is a case for open sourcing the language spec, then it's could probably be accomodated, but all the classes in J2SE? Now that would be magnitudes more difficult to accomplish, simply from a logistical point of view, never mind anything else.

    There are people who believe (and for the record I don't think the ESR is one of them) that Open Sourcing is the magic bullet for all these corporate power-struggles over the Java language and also think that %age of developer mindshare (I really hate that word!) will go through the roof, if we can have GNU J2SDK. I think that this is pretty naive, to think that there are literally millions of developers out there, who simply will not program in a given language simply because it is not part of the software libre set. It's okay for people like ESR to preach and live these ideals - they get paid for it. I on the other hand get paid to write java code, whether java is free (as a bird) or not.
  12. Win32 Arguments[ Go to top ]

    I've heard a lot of these arguments from the Win32 community.

    As for Microsoft's reasons for standardization, who knows? They didn't standardize VB. So, I don't know. We can talk all day long about how bad the evil empire is. I hear that stuff all the time. But none of that explains Sun's dogged resistance to open sourcing Java. On the one hand Sun says, 'we embrace open standards', and on the other they refuse to open source Java. Why?

    Can somebody defend the comment made by the evangelist about how Java does not qualify for open sourcing in the same way as Perl or Python because those are scripting languages (without bashing MS). That comment makes absolutely no sense.

    As for the impact on the developer and the companies who develop, maintain and ship code, it means a lot that the language is outside of the control of a single vendor. If Sun were to keel over
  13. that they should give the Java to open source. How many companies put their employers on their pay roll to just promote projects like Netbeans?

    How may other companies have community process like JCP? Com'n, we can discuss this all long, but the bottom line, who invented Java? They have every damn right to decide to give it to open source or not? Period.
  14. Win32 Arguments[ Go to top ]

    <!--Can somebody defend the comment made by the evangelist about how Java does not qualify for open sourcing in the same way as Perl or Python because those are scripting languages (without bashing MS). That comment makes absolutely no sense. -->

    You have to understand that Java is more than a language, it is more of a platform for enterprise computing than just a mere language. I am 100% agree with SUN, that you can not compare it with all other languages(don't pay too much attention on interpreted languages, I think it was just the figure of speech). In my personal opinion any one who wants Java to be open source are either insane or has a hidden agenda to Sabotage Java.
  15. Not what was said (re: What?)[ Go to top ]

    Every other major language C, C++, C# (really), Perl, Python, Ruby, etc. have implementations of their

    > compilers/interpreters in the public domain as open source. The response from Sun that this doesn't
    > apply to Java because it's not a 'scripting language' is nonsense.

    No-one has said this, please read the article more carefully. To quote PC Pro:
    > Raymond wrote: 'Sun's insistence on continuing tight control of the Java code has damaged Sun's
    > long-term interests by throttling acceptance of the language in the open-source community, ceding
    > the field (and probably the future) to scripting-language competitors like Python and Perl.'
    >
    > Phipps responded that Java is not a scripting language, so it is meaningless to make such a
    > comparison.

    Even that is reportage. What I actually said was that there will be room for both Java and scripting languages like Python and Perl and that there are even great Java-oriented scripting languages like Jython.

    S.
  16. Huh?[ Go to top ]

    How did:

    > What I actually said was that there will be room for both Java and
    > scripting languages like Python and Perl and that there are even
    > great Java-oriented scripting languages like Jython.

    Get turned into:

    > Phipps responded that Java is not a scripting language, so it is
    > meaningless to make such a comparison.

    So am I to assume that, given your revised statement, that you feel that Java is on the same level (in terms of it's ability to open source) as Perl or Python?

    Others on this list have asserted that Java is far to (complex|large|important|serious) to go open source. One went so far as to call the notion of open sourcing Java an admission of insanity. I'm curious as to what your take on that is.

    I also appreciate you coming on this list to talk about the article. Bravo.
  17. I'd've been here sooner to discuss but I was on tour and bandwidth constrained, so thanks for the welcome :-) It is always exciting getting quoted in the media. It's like listening to hip-hop; you know those samples came from the cited source but marvel at the way they are unidentifiable without attribution...

    My view is that, since JCP 2.5, new JSRs have been able to be implemented as open source with the compatibility tests funded either by the Scholarship Fund or by a generous host like ObjectWeb. The result of the JCP 2.5 rules has already been three open source J2EE projects committed to compatibility. Tiger (Java 1.5) is the first J2SE spec subject to the JCP 2.5 rules, hence is the first base Java environment that can legally be implemented as open source. Thus I do believe the Java environment is just as capable of open source implementation as any other programming system, yes.

    Now, exactly who does this is an open question. Despite the bitterness expressed by contributors to this and many other threads (and the petty and hate-filled anti-Sun campaigns to this end), I don't believe Java is solely dependent on Sun any more and neither do I believe Sun is the only party who should be pressed to make the financial sacrifice of creating an open source project for Java.


    S.
    (BTW, these are all my own opinions on my own time...)
  18. History of corporate outsourcing[ Go to top ]

    I don't think that Sun needs to 'make the financial sacrifice' for open sourcing Java. Netscape and IBM have done well be creating consortiums for their support of open source. Given Java's popularity, couldn't the same thing be done? IBM, BEA, Oracle, Sun, perhaps even Apple would financially support such a venture.
  19. Heh, heh, Freudian slip. ;-)
  20. Not just outsourcing[ Go to top ]

    While it's easy as an outside observer to just look at the eventual outcomes, Sun has enough experience of taking existing projects open source to know the issues faced (see www.sunsource.net for more). We have so far resisted the temptation of a hack (a quick'n'dirty PR job and then sort out the problems later) and I still think that's the right choice.

    We would fully expect to have to underwrite the whole activity, to have to effectively stop development for several months while we switch from internal to public processes, set up the governance and negotiate roles. To get other corporations to underwrite the venture it wuld most likely be necessary to set up a Foundation - they won't just give Sun money for 'nothing', none of those mentioned are blind altruists. Even if they would, the current engineers working on the RIs are all Sun employees that Sun would continue employing. In addition there is the matter of 'writing off' the value of the IP on our books. Keep in mind also that all the RIs feed directly into development of both JES and JDS so the impact would likely propogate.

    I could go on - I have a section in my conference keynotes on 'costs & benefits'. The 'done well' you assert is the voice of faith; Netscape's activity (in which Sun participated, you'll recall) took several years to get going and I doubt has been a net gain to either AOL or to Sun yet. IBM's activity with Eclipse still looks like a net cost to them.

    Thus the whole venture would be costly in terms of notional value surrendered, of productivity lost and of actual expenditure on staff, systems and legal advice. To date, the benefit to the Java community and to shareholders has not seemed high enough to decide to spend the time & money on that rather than, say, building Tiger.

    Going forward, I can't disclose or speculate on what Sun will choose to do but rest assured the 150,000 air miles I have travelled in 2003 attending open source events of all kinds has not been for fun, even if I have enjoyed the many discussions I have had with open source developers (yes, Fred, I have actually met some). I'd still encourage folk to consider Sun's actions as responsible rather than recalcitrant and to call for the wider Java community, not just Sun, to run with the open source baton.


    S.
  21. Two followups[ Go to top ]

    I wonder what 'notional value surrendered' means. It sounds too me like Sun wants to hang on to Java to make a better target for takeover bids.

    As to 'Java community, not just Sun, to run with the open source baton', I think you have to lead by example. At least provide a commitment to open sourcing Java and publish some sort of time frame.

    That being said, I think the Java community has been very active in open sourcing. Most of the best Java projects are open source. What are you looking for in the Java community open source effort that you find lacking today?
  22. Responses to "Two followups"[ Go to top ]

    I wonder what 'notional value surrendered' means. It sounds too me like Sun

    > wants to hang on to Java to make a better target for takeover bids.

    No need to look for 'bad' all the time, Jack. I simply mean that the intellectual property represented by the Java implementations has an asset value on Sun's accounts, and that potentially that value would need to be written off or marked down for accounting purposes.
     
    > As to 'Java community, not just Sun, to run with the open source baton', I think
    > you have to lead by example. At least provide a commitment to open sourcing Java
    >and publish some sort of time frame.

    Well, I would contend that we /have/ led by example in relation to many, many projects, but that the situation with Java is too complicated (especially with regard to maintaining compatibility) for simplistic approaches. We'll not commit to an open source reference implementation until we know we can do it.

    By the way, I'd ask people generally to talk about "an open source reference implementation" or "an open source JVM" or whatever they precisely mean - "open source Java" is not precise enough to talk about, there is already plenty of "open source Java" such as the three J2EE implementations in hand, and the huge foundry on SourceForge.
     
    > That being said, I think the Java community has been very active in open
    > sourcing. Most of the best Java projects are open source. What are you
    > looking for in the Java community open source effort that you find lacking
    > today?

    Actually I agree totally - I am very excited by and supportive of the open source Java community and have no criticisms, unlike ESR and Michael Tiemann :-) When I spoke of 'Java community' and batons earlier I was referring to the small community that creates JVMs and JREs rather than the wider open source Java developer community which Charles Ditzel points out is huge and vibrant (see http://cld.blog-city.com/read/492917.htm - it was the claim that the Java platform is somehow a place avoided by open source developers that got me so mad.

    S.
  23. My personal version of the term 'open source Java' is something similar to the dreaded scripting languages. That you could download 'everything Java' from Sun and compile and install it. Simple as that. We can do that today with Perl, Python, Ruby, GCC, the Linux operating systems, etc.
  24. I think we agree. The question, though, is not "what" but "how" - and I really don't believe there are simple sanwers to that question.

    S.
  25. Just get the source out there[ Go to top ]

    I think the first step is to just get the source out there. Make sure the files have the right license, build the distribution media and put it up on the Sun site. Viola. After that you would probably need to build a consortium to fund the Java effort and move all of the engineers into that company. Sounds like a net cost savings to me as IBM, BEA, etc. are guaranteed to fund that company.
  26. And as if on cue[ Go to top ]

    IBM urges Sun to make Java open source.
  27. What if SUN dies[ Go to top ]

    I don't really care about Java being open source per sé. I just want to be sure my investment in Java remains solid if something happens to SUN (bankruptcy, change of heart, takeover). Thje Java desktop on Linux approach seems to be akin to the Apple desktop on Unix (OSX). It is understandable that SUN wants to reap commercial benefits from Java on Linux this way. After all, Apple has also created a propietary success in a similar way. They may not succeed, but hey...

    I can also understand SUN just wanting to finally make some real money from their huge success with Java. So, I would also not be to eager to give it all away. In my mind, the core language could be open source (J2SE) and the enterprise features could be closed (J2EE). This would address the "a language should be open source" argument, while also addressing the "the enterprise stuff is the money machine and should therefore be within the realm of control" argument.

    I can't really see a real benefit for SUN controlling J2SE, apart from ensuring that certain functionality that is key to their strategy works well. The source code of Windows that was leaked provided some interesting insight here. Some customer and application specific hacks were implemented to ensure optimal compatibility with designated software. That is the kind of control that is beneficial, especially if you want to do stuff like a Java desktop.

    Marc Schipperheyn
  28. What if SUN dies[ Go to top ]

    I don't really care about Java being open source per sé. I just want to be sure my investment in Java remains solid if something happens to SUN (bankruptcy, change of heart, takeover).


    Well said Marc. This is the question I want answered as well.

    After reading the both side's opinion, and what has been said around the web, I realized that much of the facts surround this matter are't very clear.

    A few key questions we have to answer are:

    a. If something happens to Sun, to quote marc "(bankruptcy, change of heart, takeover)", is there a way for the rest of the community to make sure that Java lives on, regardless of what Sun does or does not do?

    b. Does Sun have the power to retroactively impose charges on companies that use Java? Even though the systems were commissioned while Java is available for free? Can these powers be applied to users of non-Sun Java products, such as IBM or Open Source JVMs?

    c. On the one hand, ESR suggests that Sun's licensing terms are ridiculously restrictive, while Sun says that through the JCP a lot of power has been surrendered to the community. Where between these 2 points does the truth lie?

    Ryan
  29. What if SUN dies[ Go to top ]

    I can't really see a real benefit for SUN controlling J2SE, apart from ensuring that certain functionality that is key to their strategy works well. The source code of Windows that was leaked provided some interesting insight here. Some customer and application specific hacks were implemented to ensure optimal compatibility with designated software. That is the kind of control that is beneficial, especially if you want to do stuff like a Java desktop.

    >
    > Marc Schipperheyn

    IMHO, these "hacks" are one of the reasons why Windows is so problematic, despite its recent improvements. In order to have a head over competition, MS may change some WinOS code for their Office Suite to run better than any other software in Windows, for example. They have the right to do it, being them all their own source code and all, I am not questioning it. But these hacks just increase the complexity, makes it less general and safe, and thus more buggy. I wouldn't want Java to thread this path, and maybe having Sun's implementation become Open Source would be a way to lessen the chance of these hacks to happen. But then again, we have other SDK implementations in the market to switch to, in the case these hacks started to appear, lessening the need for a Open Source Java SDK.

    By the size of Java market, the moment Sun would die, someone (IBM? Open Source community? ISO? ANSI?) would be there to keep on its evolution. Java is bigger than Sun, as someone else already said.

    By the way, who "owns" C++, is it Stroustrup still? is there a panel of specialists who decide what changes goes in C++? It was standardized by ANSI and ISO, but does it have an JCP-like panel where the industry can steer by consensus the evolution of the language? If so, wouldn't an structure similar to this (ISO standardization + industry steering) work for Java language too?

    Regards,
    Henrique Steckelberg