Analysis of Oracle's patent suit against Android's use of Java

[I D Shukhov]

http://www.eweekeurope.co.uk/comment/google-suit-now-we-know-what-java-means-to-oracle-8991

Some expressed surprise, as Sun Microsystems open-sourced Java, and issued a patent grant before Oracle bought the company, which they believe should make Java free to use. Bruce Perens, quoted the list of demands in the patent grant, saying: “It’s not clear what Oracle is after. Could it be that Google’s Java implementation does not meet the above requirements?”

It is indeed possible. Android’s creator, in an eWEEK interview last year, warned that Google’s free-and-easy approach with Java was in conflict with Sun’s Java approach. Miguel de Icaza has described the likely breakdown of relations between Sun and Google over the use of Java in Android, and the way in which the resulting situation could have been spun into a shining asset for a buyer of Sun - especially a giant, litigious one.

and

Moving away from Java?
The eventual loser from all of this will probably be Java. Despite its widespread use, there are alternatives, and now it is clear that Oracle wants to use its Java patents to get money out of companies that have big pockets, other players may be less willing to use the Java language.
Even if Oracle loses the case, Java could be seriously hurt by the protracted legal tussle.

[Peter Gibbons]

I was going to write a post about the future of Java before the lawsuit.

The conclusion was going to be:

"I don't trust Oracle to do anything good with Java."

Oracle is a sales driven organization that is trying to maximize profits.
That's great ... for them. I personally don't see any product of theirs that I would buy.

I think this lawsuit will ultimately lead Google to abandon Java for Android.
( And long term some companies will abandon Java when alternatives become available. )

What Google could do is get Scala and/or Python to generate Dalvik bytecodes directly.
Scala is already ahead technologically of the next ( unreleased ) version of Java and adoption by Google will give it a major boost.

Google is big enough and powerful technology company to have their own language(s) and VM(s).
They already have Go and V8. The Python creator works for them. They are funding the development of LuaJIT and I wouldn't be surprised if they hired Mike Pall who has achieved some amazing results with the latest version of LuaJIT - beating V8 and starting to approach the performance of the JVM. And then there is Dalvik that recently got JIT compiler.

I don't see Google paying Oracle money and be hold up by them at the same time. They may settle - but long term they will do their own thing.

[The Gorn]

Geez! Moar, moar, moar languages!!!! Reality check anyone? How stupid is this?

Can't this industry and vendors just f*cking PICK something and get to work, already!!!!??

[Peter Gibbons]

just f*cking PICK something

That something was C++ but do you honestly think C++ would have been better than Javascript for the web?
"The compiler is compiling the C++ program contained in this page. Please wait ... ... ..."

Actually thinking about this: Aussie mentioned that the western civilization runs on mainframes.
So the industry should have stopped creating new languages after FORTRAN and COBOL.

[The Gorn]

C++ - yeah, yeah...

Have some sense of proportion. There are already a jillion scripting languages, p-code type languages, etc.

Every computer scientist is indeed a special snowflake, unhappy with the status quo.

[Peter Gibbons]

Hey,

What did I say?

Google should make Python emit Dalvik bytecodes.
Python is an existing language.

Scala is a continuation of the work of one of the guys that worked on Java generics.

He made the following statement during presentation: "Hopefully this time we got them right!"

People make mistakes - why not let them fix the mistakes.

[The Gorn]

He made the following statement during presentation: "Hopefully this time we got them right!"

People make mistakes - why not let them fix the mistakes.

Because the CS field as a whole sounds like a crack addict and "let's invent a new bytecode and a new language" sounds like the payload for inhalation.

This churn - and it is churn - limits the expertise of anyone working in the industry. You cannot become very good with a given language and platform, because it is defined out of use too quickly.

C++ has an extremely rich ecosystem of best practices, books, articles, and libraries. It's been in use for >20 years and has only improved. I'm not saying that it should be used for everything, but C++ *has* improved dramatically over the years and so have practices in using it.

The C++ of 1998 - proprietary and expensive Rogue Wave C++ libraries, MFC, and inconsistent use in industry - was crap and I hated using it then. Since then, widespread practices such as the use of "smart pointers" and container classes and STL have homogenized how C++ is applied. The C++ of 2010 is a quantum leap over the C++ of the 90s.

If C++ had been obsoleted in 1998 you would not have seen this improvement.

I recall in 1995 that Java was heralded as a millennial advancement. At the time my instincts were that it was proprietary so I wondered just how far it could go. Now it is yesterday's trash. I know it's legally tied up by a shitty company (Oracle) but the fact remains, an argument for "new language" can always be made.

Summary - I think the impact of heralding new computer languages has been trivialized. I don't like forced obsolescence because it marginalizes the human capital, namely the skills of those who are invested in current tools. Ultimately this sort of thing keeps this industry immature and "de-skilled."

New computer languages are proposed and rolled out these days the way major libraries used to be a couple of decades ago.

[lorb]

If not Java in some form, then C# would seem to be the choice.  The thing is, I don't know how good the de Icaza version of C# really is, even if it is a version or two behind MS's .net.

[JavaMouse]

Well, I agree with G0ddard in thinking that new languages are cluttering the landscape insanely nowadays.  I understand that CS majors need something to do.  But does it make sense to do real development in these languages? I dunno.  It looks like it's just happening whether you want it to or not.

I think this lawsuit will ultimately lead Google to abandon Java for Android.
( And long term some companies will abandon Java when alternatives become available. )

What Google could do is get Scala and/or Python to generate Dalvik bytecodes directly.
Scala is already ahead technologically of the next ( unreleased ) version of Java and adoption by Google will give it a major boost.

Google is big enough and powerful technology company to have their own language(s) and VM(s).

I don't really grok the lawsuit, so I don't know whether this is a serious problem for "Java" per se or not. Isn't the "problem" that Google was doing something iffy with Java, and Oracle wants them to lay off?  So far as I know, most companies do not mess with the JVM, they use it as is.  So why should anyone else worry?

Regardless of that... I started looking into Scala.  It's interesting.  I've also been doing more work with Javascript.  Doesn't hurt to keep your options open.  Besides, all the k3wl kids are doing it.

[lorb]

So far as I know, most companies do not mess with the JVM, they use it as is.  So why should anyone else worry?

Not only that, but think about what we are saying here for a second, Google can't afford to pay a license fee from Oracle?  If that isn't the most ridiculous thing.  Why should I take Google seriously if it's getting perceived as a "we could only afford to do it out of our garage" act?

Just because Google is a SW co., they have to invent their own browser, and have their own OS, and "own" their own language?  To me that is the idea of developing own SW and being able to use it free just because one is a SW company.  Lots of time, effort, and money (even if out-of-pocket) was poured into these technologies.  Why should I look at Google like "Cool, they are scoping out some FOSS code from The Man".  They are "The Man", why should I feel sorry for them?

To me, the way you go about it is get a 20 year license, that way can benefit from the huge early adoption phase and after that whatever is going on will be so ubiquitous that the price of "What the market can bear" will be much lower, and you have gained the early marketshare.  But I would expect the classic ubergeek to not get how business works and shout "It's not free!"  The real threat would then be said licensor trying to renogotiate later on and say "that's not in the contract", but that's life, you pay to play.

[JavaMouse]

Just because Google is a SW co., they have to invent their own browser, and have their own OS, and "own" their own language?  To me that is the idea of developing own SW and being able to use it free just because one is a SW company.

There are two "problems": Google is awash in money and is trying to find ways to do keep the bloated stock price high.  Their MO seems to be to try to buy or incubate numerous internal startups and hope that some hit the big-time.  In addition, as mentioned in Paul Graham's recent inflammatory essay (which basically said Yahoo's coders suck), Google has a "hacker" culture in which impractical, ivory tower pursuits are encouraged.  And they can afford to let that run wild because they are awash in money.

They probably do have a story about how their own OS, language, etc are going to keep them at the top of the heap.  Whether or not it comes true, we'll see.

Given the culture at Google I'm surprised they didn't build something that was truly their own for Android, that would not have been open to a lawsuit.  Presumably the Google lawyers have already decided that Oracle does not have a case.  I'll take a wild guess, but I bet Google's lawyers are more nasty and high-powered than Oracle's, so I'm betting on them to win, but again we'll see.

[lorb]

Thanks for the insight Java-mouse.   

Only part where I would differ is I would bet a judge or jury would be able to see it from Oracle's point of view - we are talking about regular people unless they get a jury full of geeks.  Maybe they will comb "Slashdot" for jurors.   

Okay, I read that PG article.  Pshh, how can a company be hacker/startup-like when they have more money than they know what to do with.  That is like saying "Go play off in the corner and hopefully something good will come of it."  I could see that in the days when (the idea of) patents were still king, but it's hard to think that makes a ton of sense these days, or if were "standards bodies" working on a new protocol, then perhaps that would make some amount of sense, not just "Go work on the 'Google standard'" for whatever.  They can because of cash, but I disagree with the kudo Paul seems to be attributing to it in order to boost the "hacker culture" mantra.

[The Original Henry]

Google has a "hacker" culture in which impractical, ivory tower pursuits are encouraged.

I think the rest of the world calls this "research and development". Google is one of the rare software companies with both enough capital to pay for it and the dedication to actually do it on a regular basis.

[Peter Gibbons]

I recall in 1995 that Java was heralded as a millennial advancement.

Sun did great marketing for Java in the early years.

Now it is yesterday's trash.

There is a Java language and then there is the Java VM.
The consensus seems to be that the Java language is starting to show it's age, however the JVM has improved a lot over time and it's fast and stable.

That's why there are new ( and not so new ) languages that target the JVM. Scala, Clojure, JRuby, Jython.

Google has done something different - they replaced the JVM with their Dalvik VM but kept the Java language.

I don't see any problem with both approaches. I fact I didn't see any problem with Microsoft cloning Java and it's class libraries and creating C# and .NET.

Microsoft and Google are businesses - it's a very sound business decision not to depend on your competitors' technology.

The only place where you could have monoculture type of development is the government. They could mandate that everybody uses C++ or Ada and you have to comply.

[lorb]

Microsoft and Google are businesses - it's a very sound business decision not to depend on your competitors' technology.

How are Google and Oracle competitors, though?

I never understood the MS vs Google competition (brain-drain, give me a break) other than in a sense of MS wanting to rule the SW world and Google poaching employees from MS (so what? equilibrium is bound to happen).

Nice explanation of the tech-side.