Day 3 of Oracle v. Google: Larry Page bests David Boies | Hacker News Reader

Day 3 of Oracle v. Google: Larry Page bests David Boies

groklaw.net

77 pointsgrellas26 comments

Strong play by Google's lawyers here.

    Google: This means that a JME app may not run on a JavaCard device?
    Edward Serevan: Yes.
    Google: There are different APIs between JavaCard and JSE?
    Edward Serevan: [Some disagreement over the differences between versions.]
    Google: There are differences between JSE, JEE, JME, and Java Card?
    Edward Serevan: Yes.
    Google: How can it be that (Oracle can say that Java is WORA)?
    Edward Serevan: Oracle defines WORA as across platforms [different operating systems], not Java versions.

I'm not sure how Oracle is going to get the assertion that Google caused significant damage to the Java ecosystem by their violation of WORA to stick here, in light of this testimony. By any objective measure, Google has created a new platform for Dalvik in Android.

They are going to assert the Android is for phones, and J2ME is for phones. Thus Android broke their WORA definition for that platform. Seems fairly straightforward and that Google is kind of making their case for them.

(Just to clarify: I think Oracle should, and is likely to lose their case.)

I think that's easily trumped, though, because "phones" are not a platform. They're a form factor. Otherwise, J2ME should run on iPhones and Windows 7 devices.

Part of the whole Android package is a hardware specification [1]. Android The Product is a platform, which consists of a compatible hardware device and a Dalvik VM with Android APIs. It isn't just the software. J2ME might happen to run on some handheld battery-powered computers, but a form factor does not a platform make.

[1] http://static.googleusercontent.com/external_content/untrust...

I don't think that saying J2ME should run on any given OS for a platform or form factor is relevant. I could then argue that if there is no Java for Minix or someone's pet OS that runs on a server somewhere that servers are not a platform for Java. Doesn't seem right.

I don't think it matters how complex or precise the Google definition of what Android is, that is their definition. Oracle's point in this was the damage to the perception of Java as a portable way to build programs for anything "Java" -- Android was only part-way Java, creating the alleged confusion. I think the parallel to Sun's MS suit from long ago is fairly straightforward.

     I think the parallel to Sun's MS suit from long 
     ago is fairly straightforward.

Not it isn't.

Microsoft had Java licensed from Sun, meaning that they could use the Java trademark for their own compiler and their own VM. In return Microsoft had to make their compiler and their VM be compliant with Java's specifications.

Microsoft then started to extend everything, the language, the base class library (and you cannot extend the base namespaces, e.g. java.*) and their virtual machine. They did this while advertising to developers that their compiler takes Java code and produces Java-compatible bytecode.

The differences:

     - Google never agreed to an (explicit) license
     - Dalvik is not a Java VM and was never advertised as such
     - Google does not claim Android has Java technology inside
     - Google does not claim that the compiled binaries can be
       executed on Java VMs

What Google actually did:

     - created a compiler that translates from JVM bytecode 
       to Dalvik bytecode
     - took the base Java library from Harmony, a 
       clean-room implementation

What Google did is exactly what projects like IKVM.net are doing. IKVM is a project that allows you to recompile and run Java jars on top of a .NET VM.

This is not new and if Oracle wins it will open a huge can of worms, as if our industry wasn't having enough problems related to "intellectual property" already.

Google did not agree to a license but if Oracle convinces the court they needed one then that is a moot point (Oracle asserts trademark, copyright, and patents to make this claim I believe). All the differences between what MS and Google did are irrelevant if Oracle can convince the court that having Java all over the developer website and documentation created confusion with respect to what Java is.

I think to a relatively non-technical judge they have a chance -- I can imagine a judge asking himself/herself, "So you used the language Java and the libraries for Java but you said that it wasn't Java. But if it walks and talks like a duck and you call it a chicken are you still not creating confusion for anyone who knew what a duck was before you showed them your 'chicken'?"

I agree that this would line projects like IKVM up to get nailed but that has nothing to do with the merits of the case.

Also: "Edward Serevan: I wouldn't characterize the JSE as being only for laptops and desktops. Smartphones have become capable enough to run the JSE."

It looks like their expectation was that JSE would run on phones as well. So they were already planning fragmentation at this level.

Agreed, but Android was never actually J2SE either so Android is a fragmentation of a different sort. Of course, the the court accepts Oracle's right to control Java then it is Oracle's choice to fragment it however they like, not Google's. If Google is out ahead of them defining how this fragmentation will happen then if you are Oracle I guess you argue that this is damaging their ability to plan and execute it as they wish since Google's choices are already associated with the Java name. This confusion is, I guess, sort of, kind of, damaging to the ecosystem.. if you were easily confused I guess.

I'm not sure what is being contested in the case any more. Larry Page testified that their implementation of Java was a clean room implementation. Is Oracle trying to assert that an API is something that is subject to copyright?

The trial is examining both copyright and patent issues. But yes, one of the things Oracle is claiming is that they can copyright their API and you can't reimplement it without a license.

This does not seem like it should fly, in my inexpert opinion.

If it does fly, Oracle is going to have a lot of fun.

For instance, Oracle copied SQL from IBM years before it was a standard. There's also Sun's involvement in Wabi. I'm sure Microsoft wouldn't mind extracting some cash for that "infringement".

Novell and the Unix API...

It shouldn't fly. We already went through this with IBM's BIOS and its clones in the 80's.

We've been discussing this extensively in other threads: http://news.ycombinator.com/item?id=3862782 http://news.ycombinator.com/item?id=3859361 http://news.ycombinator.com/item?id=3854767

But the answer is yes, they try to say an api is subject to coypright (in the article is a list of challenged points, including apis).

I enjoy reading both the incredibly biased but still informative Groklaw and then the incredibly biased in the opposite direction (and paid by MS/Oracle) Florian Mueller. Looking past the thinly veiled cheerleading isn't that hard once you realize where they're coming from and both seem fairly thorough.

But what would be really awesome is if there was somewhere I could go for legal tech blogging that isn't always good or bad news for open source.

Nilay Patel and Matt Macari at the Verge do a good job but coverage isn't as regular as what you'll get by checking Groklaw daily and it gets lost in the Verge-o-polis.

I've always found the Nilay Patel legal coverage, despite his qualifications, to be overwhelmingly tainted by whatever the Apple fanbase would currently like to be legally true. Like Florian I think he has a direct monetary incentive for taking a certain view and presenting it as objective.

I mean this honestly: what makes you think Groklaw is biased? That deosn't match my [uninformed] mental model of their site.

I don't think PJ ever hid the fact that she was advocating in a certain direction, quoth wiki:

"Groklaw publishes articles (both news and opinion) from a self-described pro-FOSS, anti-FUD perspective"

The comments are similarly policed to comply with this. This certainly isn't a secret, as Florian Mueller's deals with MS and Oracle were, and I don't think I'm making a wild accusation here.

Back in the day a 100% anti-SCO line worked great because SCO was a fraud. But the current issues are murkier and the one sidedness of the site now sticks out like the news channel at your grandparents house.

Misspells the Oracle software architect Edward Screven as Edward Serevan throughout.

Oracle mobile phone plans revealed

http://ofb.net/~pcullen/ophone/ophone.html

Gonna start to get pricey with multi-core phones.

How strange is it that Eric Schmidt guided the development of the very software Google is now accused of copying?

What is SSO?

I just searched the linked page and found this: SSO (Sequence, Structure, Organization).

Thanks.

Reply on news.ycombinator.com