Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Google Java Oracle

Oracle's Newest Move To Undermine Android 342

GMGruman writes "Oracle's decision to shift focus from the Harmony Java open source project to OpenJDK seems innocuous enough — but InfoWorld's Josh Fruhlinger explains it's part of an effort to derail Google's mobile Android OS by gutting the open source project that Android has been driven by. IBM has signed on, apparently in return for getting the Java Community Process reactivated, leaving Google in a bind."
This discussion has been archived. No new comments can be posted.

Oracle's Newest Move To Undermine Android

Comments Filter:
  • by WrongSizeGlass ( 838941 ) on Tuesday October 12, 2010 @03:32PM (#33874898)
    Google is full of smart people. I'm sure they saw this move - and the entire assault on mobile Java and derivatives thereof - coming long before Oracle started their anti-Android crusade. I'd be willing to bet that Google has something new 'brewing' for Android 3 that will leave this whole mess behind. You just don't get that many programmers together without a few being paranoid enough to have planned an 'escape module'.
    • Re: (Score:3, Insightful)

      by Orga ( 1720130 )
      Agreed, I think the usage of the term bind is excessive. Somehow MS and RIM survive without external developers working on their products language I'm sure Google can handle it. And as TFA stated Google has OpenJDK developers right now, whose to say what the future will bring to android.
      • Agreed, I think the usage of the term bind is excessive. Somehow MS and RIM survive without external developers working on their products language

        What language does RIM use again?

        Oh right, Java [blackberry.com].

        As for MS, both C# and the Common Language Runtime have published specifications, and MS's implementation is not the only one (see also: Mono Project). Granted, I don't think Mono has released a version of .NET Compact Edition.

        • .NET CE is an implementation of the same CLI spec, just for a different platform. I'm not sure if Mono runs on WinCE, but it does run on Linux/ARM, which would count as an analog.

    • Oh god, you scared me for a second when putting "brew [wikipedia.org]" and Android in the same sentence.

    • by alen ( 225700 )

      the problem is not lack of devs, but money. Google currently has some good margins and takes in a lot of cash. if they were to develop their own visual studio clone for android development it would cost a lot of money that they would have to eat in lower margins and lower stock price. or charge money for the kit making the cost of entry a lot higher than it is now. kind of like apple requiring the purchase of a Mac to develop for the iPhone

    • Don't taunt the Google.
    • by Fnkmaster ( 89084 ) on Tuesday October 12, 2010 @04:07PM (#33875416)

      Not even check, honestly. More like a limp attack on the queen with a pawn. What Google uses from the Harmony project is a bunch of the core java.* classes. This stuff changes, sure, but not particularly heavily or rapidly these days. This is not where Android is innovating, nor is it a huge area of rapid development, assuming Harmony is at or approaching stability. This might require Google to shift a couple of their Java developers around, but the legal issues are far more significant than any costs associated with this.

      The Dalvik VM itself is already developed internally at Google. The Android apps and framework and the rest of the stack is already developed internally at Google.

      This might very well mean that Harmony won't see ongoing development toward being a fully featured JDK replacement, but Google doesn't need that anyway.

      I'm not an expert on Android internals or anything, but I think this story is being significantly overblown.

    • I'd be willing to bet that Google has something new 'brewing' for Android 3 that will leave this whole mess behind.

      But what would the brew be, that would work with all of the existing applications written in Java today?

      You mention programmers being paranoid which implies a technical solution, but how can even the smartest programmer have developed an escape module from what is essentially a legal problem...

      You also imply some Java variant to be switched to ("brewing" in quotes) but that doesn't get around t

      • But what would the brew be, that would work with all of the existing applications written in Java today?

        It would be the Brew that is True, so Google wants to avoid the Vessel with the Pestle. I'm sure this means they'll choose the Chalice from the Palace.

  • by BobMcD ( 601576 ) on Tuesday October 12, 2010 @03:33PM (#33874906)

    Forgive the layman here, but why can't Android simply switch Java platforms as well? Open is Open, no?

    • by WrongSizeGlass ( 838941 ) on Tuesday October 12, 2010 @03:37PM (#33874968)

      Forgive the layman here, but why can't Android simply switch Java platforms as well? Open is Open, no?

      Oracle is trying to claim that Dalvik, Android's virtual machine [wikipedia.org] infringes on mobile java patents. Mobile java was not included when Java received it's current "open" licensing.

      • by Kostya ( 1146 ) on Tuesday October 12, 2010 @03:55PM (#33875230) Homepage Journal

        Oracle is trying to claim that Dalvik, Android's virtual machine infringes on mobile java patents. Mobile java was not included when Java received it's current "open" licensing.

        And I'm sure part of the reason why Mobile Java wasn't in the "open licensing" was the carriers. That is, Sun had already extracted some money out of the carriers and met with a very nice bit of success there. Remember, before Apple's iPhone and Google's Android, JavaME was a big success in offering advanced features (that sucks compared to today's offerings). It was a big success for Sun licensing wise--something the original Java was not.

        But with that money came a very, very hefty price. They had to bend over backwards to give the carriers what they wanted in order to "add value". One of those was charging developers $500+ a pop to be able to release applications for their network. Another for the developers to pay extra to access certain features (location). And another still was for companies like Verizon and Sprint to just flat out turn off certain features.

        Which is why Apple didn't do JavaME (I remember being pretty bummed when they didn't)--they wanted complete control, and they would never get that with JavaME.

        And Google had similar needs--but also didn't want to pay the licensing costs everyone else did.

        JavaME was a money maker for Sun (unlike the standard Java VM), but the process of making money off of it made it a nightmare to deploy apps on. Development--writing code--was ok, but getting it to work on multiple headsets (nevermind multiple carriers) was a huge headache. And it was a huge headache because of all the compromises Sun made to get the carriers on board. And that nightmare (in addition to licensing costs) is why Google came up with their own VM implementation.

        I used to be a big Java proponent for mobile development. I'm not anymore. But it is interesting to see how all those bad decisions (I cursed Sun weekly as I tried to wrestle another carrier or headset down) played out into what we have now.

        Google didn't want to pay the money. Microsoft (via Miguel) likes to say they would have been better, but they are just as bad on the licensing (see HTC and now Motorola). Sounds to me like Google got used to their free ride on Java and balked at the idea of giving anyone a slice of their work and money on Android.

        I'm not saying Ellison is not squeezing them (he definitely is), just that Google is kind of getting a bucket of cold water in their face about how the tech companies "collaborate" in new tech fields. Not "fair", but it is kind of predictable.

        • Re: (Score:3, Informative)

          by stephanruby ( 542433 )

          And that nightmare (in addition to licensing costs) is why Google came up with their own VM implementation.

          Technically speaking, RIM (Blackberry) was the first mobile OS to come up with that workaround. Google (or more specifically the startup that Google purchased) just saw what RIM did, saw that it was working, and did the same themselves (thus avoiding the per JVM licensing fee that Sun was charging companies).

      • by Homburg ( 213427 )

        But Oracle is claiming that these patents apply to whatever Android is doing now, using Harmony, so I don't see how it would make any difference if Android switched to using OpenJDK instead, as the OP suggests.

    • by mcgrew ( 92797 ) *

      I RTFA, and I don't see how it will hurt Google at all.

      The components of Android that allow it to run Java code are based on the Harmony project, an open source implementation of Java created under the aegis of the Apache Software Foundation.

      This is a non-issue, unless I'm missing something. The writer's point is that Google will have to contribute more to the project since IBM had done so much of it, but so what? As someone mentioned earlier, Google has plenty of talent to throw at it.

    • by Homburg ( 213427 )

      Well, Harmony is released under the Apache license, while OpenJDK is released under the GPL. I don't know if that would make a difference - Android already includes GPLed software, so Google clearly have no objection to the license in general.

  • Who is surprised (Score:5, Insightful)

    by h4rr4r ( 612664 ) on Tuesday October 12, 2010 @03:33PM (#33874912)

    Did this surprise anyone?
    Let us all remember that ORACLE stands for "One Rich Asshole Called Larry Ellison"

    This is the company that buys out someone else and does not even bother to offer the customers a migration path. Nor any form of support other than letting you fill out a bug report they close as the product is EOL.

  • Given that this project is only just starting, why can't they just port everything that they need from Harmony into OpenJDK and change over in V4?

    I know this is simplictic, but as an idea, where would it fail?

    • Sorry, project is the wrong word, I should have read it one more time before hitting post. Given that this fight is only just starting....

    • Given that this project is only just starting, why can't they just port everything that they need from Harmony into OpenJDK and change over in V4?

      At a guess, licensing. OpenJDK uses the GPLv2 (GPLv2+Classpath exception for some files). The Apache 2.0 license isn't compatible with GPLv2, only GPLv3.

      The Classpath exception covers the linking of any files covered by it in other software projects without requiring said projects becoming GPL.

  • by Anonymous Coward on Tuesday October 12, 2010 @03:41PM (#33875028)

    avoid the fees B.S. and just ship the 100 meg java SDK with android and be done with it. it even has a patent cross licensing clause. yes its bloated. yes developers might not use any of its features. who the fuck cares ? just ship the damn thing and keep the JVM compatible. if a nokia dumbphone from 5 years ago can ship with j2me so can an android smartphone.

  • No Duh. (Score:5, Insightful)

    by Bill_the_Engineer ( 772575 ) on Tuesday October 12, 2010 @03:42PM (#33875042)

    Sun started OpenJDK as the project from which the GPL'ed version of Java would be created.

    It stands to reason, that Sun had planned to discontinue supporting Harmony when OpenJDK was formed.

    Don't mean to spoil a good conspiracy...

    • Re:No Duh. (Score:4, Insightful)

      by dr.newton ( 648217 ) on Tuesday October 12, 2010 @03:47PM (#33875118) Homepage

      Did Sun ever really support Harmony?

      Either way, making a deal with another company to ensure that all their developers stop working on a project is going farther than to "discontinue supporting" it.

      Also, I think you did mean to spoil a good conspiracy. Shame on you.

      • Re: (Score:3, Interesting)

        Did Sun ever really support Harmony?

        Actually no. I don't think so. But the summary gave the appearance, so I just gave it some latitude.

        Either way, making a deal with another company to ensure that all their developers stop working on a project is going farther than to "discontinue supporting" it.

        Except that IBM has said that they have no plans to stop supporting Harmony. Of course the exact words were " IBM will continue working on Harmony, but its main efforts will be directed toward OpenJDK, Smith said.

      • Re:No Duh. (Score:4, Interesting)

        by DragonWriter ( 970822 ) on Tuesday October 12, 2010 @04:29PM (#33875722)

        Did Sun ever really support Harmony?

        No, IBM did.

        This article is largely about Oracle offering IBM concessions regarding the management JCP process so that IBM would drop involvement with Harmony in favor of dedicating resources to OpenJDK and the Java Community Process.

    • Also if Oracle (who now owns Sun) decides to throw their weight behind OpenJDK and not Harmony, can you really blame IBM for abandoning Harmony in favor of OpenJDK?
  • by digitaldc ( 879047 ) * on Tuesday October 12, 2010 @03:42PM (#33875044)
    Well I was going to read about this, but all of a sudden some Xerox jerk comes along and spills papers everywhere blocking the text of the article.

    I got so annoyed I just left...thanks InfoWorld/Xerox!
    • Re: (Score:3, Funny)

      by mcgrew ( 92797 ) *

      They must not like your browser, it didn't do that to me. But you didn't miss much -- TFA's author was really reaching, unless I missed something.

  • by shutdown -p now ( 807394 ) on Tuesday October 12, 2010 @03:43PM (#33875060) Journal

    We already know that Dalvik VM itself isn't like JVM. It can be mapped one-to-one (at least going from JVM bytecode to Dalvik bytecode), but the basic architecture is different.

    Android also has its own rich class library, while retaining some stock fundamental Java classes. Of those some are inherently implemented mostly by the VM (Object, String...), so presumably they are also Dalvik-specific, while others have Java implementation - collections, for example. I assume the latter is what is taken from Harmony. The obvious question, then, is - how much code is that? Somehow, I suspect that it's not all that big, and so Google could just take over those bits it needs - rather than Harmony as a whole - without having to contribute significant resources to it.

    • Android also has its own rich class library, while retaining some stock fundamental Java classes.

      Those "stock fundamental classes" were either taken or adapted from Apache Harmony, which is the point of this article.

      • That's what I wrote. The question is, how much code is that in practice? TFA says that maintaining it would be a significant burden, but I'm unconvinced. I suspect it's a fraction of the Google-specific Android Java code in practice.

    • Of those some are inherently implemented mostly by the VM (Object, String...), so presumably they are also Dalvik-specific, while others have Java implementation - collections, for example. I assume the latter is what is taken from Harmony. The obvious question, then, is - how much code is that?

      It's a huge amount of code when you consider how reliant Java code is to how strings work, to how the networking classes work, to how date handling works, to how internationalization works...

      At the core of any modern

      • Re: (Score:3, Insightful)

        It's a huge amount of code when you consider how reliant Java code is to how strings work, to how the networking classes work, to how date handling works, to how internationalization works.

        Is it a huge amount of code, or a huge amount of work to get it right?

        Note that there is no talk about rewriting anything. Harmony already has the implementation, and it is OSS, and will remain such even if its development is discontinued. Google would only need to fork that code and maintain it on the level of fixing bugs.

        • Is it a huge amount of code, or a huge amount of work to get it right?

          Both.

          Note that there is no talk about rewriting anything. Harmony already has the implementation, and it is OSS, and will remain such even if its development is discontinued.

          Irrelevant, they already use Harmony. It doesn't matter how open it is when you can't use it because of patent issues, which is Google's current problem. That's when you start talking about re-writing code to work around patents.

          That's less work, but not at all ins

          • Irrelevant, they already use Harmony. It doesn't matter how open it is when you can't use it because of patent issues, which is Google's current problem. That's when you start talking about re-writing code to work around patents.

            Oracle's patents don't have anything to do with Harmony. They mostly relate to implementation techniques for VMs (not just Java VMs, but VMs in general - there were claims around that Microsoft licensed those patents from Sun back in the day so as to not get sued for .NET), and target Dalvik, not Android class library.

            TFA not about Oracle threatening Harmony with patents, either. It's about Oracle withdrawing official support and contribution to Harmony, so that it becomes "just a community project", which

  • What about GNU JRE? I know ORACLE owns Java when they bought Sun, but is that only Sun's implementation, or can they charge you for using free implementations, too (via patents I assume)?

  • by alen ( 225700 ) on Tuesday October 12, 2010 @03:47PM (#33875122)

    Google is famous for building a piece of cool software to version .8 or so and then releasing it under open source and letting everyone else finish the work. they build some cool software for internal use but for all their consumer products they expect everyeone else to finisht the work or let a cool product like google reader languish

    • UMADBRO?
    • by rotide ( 1015173 )
      Would you prefer the alternative of just not releasing it as open source and letting it die quietly instead? But then people would say "why can't they just release the source code and let the community run with it?". Guess you just can't win...
    • by John Whitley ( 6067 ) on Tuesday October 12, 2010 @04:23PM (#33875642) Homepage

      Google is famous for building a piece of cool software to version .8 or so and then releasing it under open source and letting everyone else finish the work.

      I call your bluff: Show source control logs that demonstrate that any significant Google open source release (of which there are many) has more than a trivial percentage of non-Google contributions. For full credit, you must show that these non-Google contributors were somehow not working in their self interest by contributing to the project.

      On that latter point... Last I checked, "open source volunteer sweatshop" was still equal to the empty set. I.e. no one is forced to contribute to any particular piece of open source code. The deal for all OSS projects is essentially the same: "hey, I made something cool, come help out if you like!" Whether "I" is a corporation or one or more independent volunteers is irrelevant. Any external contributors to a project do so for their own reasons, reasons which have been extensively discussed elsewhere.

    • by Tanman ( 90298 )

      No, google is famous for building pieces of software to .8 or so, releasing them to beta, and then letting them die.

      There are, of course, some notable exceptions. But I don't see very many google labs products get picked up once Google smurfs them.

  • by Snap E Tom ( 128447 ) on Tuesday October 12, 2010 @03:59PM (#33875300)

    And Larry Ellison's good buddies with Steve Jobs. Coincidence? I think not.

    • by Mongoose Disciple ( 722373 ) on Tuesday October 12, 2010 @04:22PM (#33875634)

      And Larry Ellison's good buddies with Steve Jobs. Coincidence? I think not.

      Of course it's not a coincidence. Pompous douchebags like other pompous douchebags. :)

      Not to derail the conspiracy angle, but sometimes it's better to bet on people being self-promoting jerks than people being Evil with a capital E.

    • Eric Schmidt was on Apple's board. So what?

      The whole case is about cashing in on the success of Android. Oracle sells dull clunky buggy database software and they'd love to get some revenue from the consumer market.

  • A Chess match. (Score:4, Interesting)

    by Roskolnikov ( 68772 ) on Tuesday October 12, 2010 @04:05PM (#33875376)

    Sergey vs. Larry.

  • Seriously. If I was Google, I would block all of Oracle's IP space from accessing any Google service. And, hell, IBM's too, for good measure. Then I would threaten to block the IPs of all of Oracle's big customers.

    What is Oracle going to do? Google isn't a public utility. Yeah, Oracle would obviously sue Google immediately, but that would work in Google's favor- they could say "Open up Java, and we'll unblock you."

    For that matter, what law says that Google has to give you access to their site? I can't think

    • Seriously. If I was Google, I would block all of Oracle's IP space from accessing any Google service. And, hell, IBM's too, for good measure. Then I would threaten to block the IPs of all of Oracle's big customers.

      Oracle and IBM have a lot of pull with most Fortune 500 companies. Starting a war with them would be a serious mistake, particularly for a company whose major source of income is advertisements.

  • JVM's have a fixed cost to develop them, just as Dalvik has a fixed cost to develop new and exciting features like JIT's and better GC's. IBM has J9, BEA (now Oracle) has JRocket, Sun (now Oracle) has Hotspot. Instead of writing their own customized JVM's they've decided to collaborate (I assume) to cut costs and streamline the JVM's presentation and roll-out to customers. How is this some super secret attack on Android?

    Harmony is open and will remain open for the foreseeable future. Its not like BEA/Oracle

    • Harmony is open and will remain open for the foreseeable future. Its not like BEA/Oracle/IBM were big supporters of harmony before this deal was inked...

      Actually, you'll find in the article that most of the code in Harmony was written by IBM employees on company time, and now that IBM is switching its efforts to OpenJDK, the ASF folks are essentially saying that Harmony is dead. While we'd all like to believe that open source projects magically maintain themselves through the gumption of citizen-programme

  • It's really Google that undermines Android by using the Dalvik VM, with its (even if slightly) incompatible bytecode instead of actual 100% Java that can run on any properly compatible JVM. That move just opens the platform to this kind of disruption. Why did Google do it, anyway?

    If Google had made Android simply a (perhaps heavily) patched Linux kernel distro, with its own variation on the GNU tools and userland, with a standard JVM, it would have tapped the entire large and dynamic Linux developer communi

    • Re: (Score:3, Interesting)

      by codepunk ( 167897 )

      No it should never have had a VM in it there is no need for it. Apple got it right by going 100% native, a jvm on a phone serves no purpose other than to burn precious processor cycles and battery life.

  • The only free mobile phone stack out there, from top to bottom, is MeeGo (formerly Moblin and Maemo). It is on Linux (GPL), then on Qt (GPL). There are NO FEES except for being in the Nokia Ovi (App) store. The problem is then handsets... or is it? Qt branches work on Android (NDK, marginally using Java via a stub loader) and in iPhone. So If you want to side step the license issue, and get platform independence, then Qt is the way to go.

  • by AC-x ( 735297 ) on Tuesday October 12, 2010 @06:52PM (#33877268)

    But with no major financial backing for the development of its Java libraries, Android could slip behind and lose the love of its Java-savvy developer base.

    Doesn't Google count as a major financial backer?

It is easier to write an incorrect program than understand a correct one.

Working...