Oracle Thinks Google Owes $6.1 Billion In Damages

An anonymous reader writes "When Oracle acquired Sun in 2009, the company got its hands on a lot of desirable technology. While OpenOffice may have fallen by the wayside, Oracle isn't about to let the Java programming language and its associated patents remain untouched if they can generate some additional revenue. In fact, the company is currently in the middle of a legal battle with Google over those patents that could potentially net Oracle billions and leave Android crippled. In August last year Oracle sued Google for infringing Java patents and copyright by developing Android. Oracle argues that Android uses technology derived from Java and therefore infringes multiple patents. It wants compensation, but with most court documents and details not publicly available, it's hard to know specifics. However, new documents made available late last week revealed just how much Oracle thinks is an acceptable damages payment for Google to make. According to an expert Oracle hired, Google could be looking at a bill between $1.4 billion and $6.1 billion for its alleged infringements."

Comments

Yesterday?

[Anonymous Coward]

Wasn't this posted yesterday?

Re:Yesterday?

[Anonymous Coward]

All my troubles seemed so far away....

Re:Yesterday?

[Anonymous Coward]

If Google had any brains they'd just invent their own stuff. They have the manpower. Why even fuck around with piece-of-shit JAVA / Linux anyway? Write some brand new low-overhead OS that is blazingly resourceful and not so full of legacy SHIT. Develop your own badass language that takes the best of C# / JAVA / RUBY and make it fucking SCREAM....... FUCK......we all want it.

Re:Yesterday?

[Anonymous Coward]

http://golang.org/

Linux is fine, Java is the problem. And they're working on a better language already.

Re:Yesterday?

[icebraining]

Developers developers developers.

Re:Yesterday?

[Hylandr]

Child...

- Dan.

Re:Yesterday?

[Anonymous Coward]

Now litigation's here to stay...

Re:Yesterday?

[Anonymous Coward]

And build their new language on top of say, a VM? Patents exist on language concepts, on dev tools, on OS's, on hardware etc. Their competitors, Microsoft, Oracle, Apple, etc have a shitload of patents in each and every area.

Re:Yesterday?

[shutdown -p now]

Until Go gets some kind of generics for libraries to provide generified algorithms and data structures, it's not worth the bother.

Re:Yesterday?

[nukenerd]

Even if they wrote a new OS or language, patent trolls would still come after them with claims of infringement.

Some patents so broad and vague that the trolls could claim that you infringe them just by getting out of bed in the morning. They do not depend on coding details. With such patents, they will go after anyone with money.

Compensation for Java?

[TheLink]

Hey maybe Oracle should be paying some of us damages for inflicting Java on the world. ;)

Re:Compensation for Java?

[Smallpond]

Offset by saving some of us from C++.

Re:Compensation for Java?

[poetmatt]

Why can't oracle pay damages for inflicting itself upon the world?

Meanwhile, they haven't even gotten to the point of reduced claims in court, so this could be cut by 80, 90, 100% by the time the claims reach the court. Hell, this hasn't even gone to trial, and google has signaled clear intentions to defend themselves in this case.

Re:Compensation for Java?

[Anonymous Coward]

Or, saving the world from you writing C++.

Re:Compensation for Java?

[kimvette]

Are you serious? Are you somehow implying that Oracle isn't in the right?

$6.1 billion in damages? Who the hell are they trying to fool? What did they lose out on: a sizable market share of free?

Oracle suffered no damages whatsoever; most of Java is licensed under the GPL (http://en.wikipedia.org/wiki/Java_%28programming_language%29) so even though Google has reimplemented a lot of the functions and basically created a Java clone, Oracle has suffered no damage because it is software they GIVE AWAY FOR FREE.

Now, before you jump on me and correct me by saying it's a patent issue, I understand the distinction. However they gave away the software under the GPL and the GPL stipulates that the software can only be distributed without being encumbered by patents, so either Oracle is right and thus is in violation of it's (Sun's) own terms for Java, and Java needs to stop being redistributed (not an entirely bad idea) or Oracle is in the wrong, and Google has done no wrong by reimplementing the Java language. In any event. Oracle is full of shit; they have suffered no damages whatsoever.

I know Oracle hates free and their RDBMS licensing fees are completely ridiculous (licensed per core x RAM - they don't care if it's a server or if you need a seat for a development or QA lab workstation they license it based on what a given CPU "could" theoretically handle) but they fucked up; if they hate free software they should not have purchased Sun in the first place. Sun's processors were "open source," their office suite was, they opened up most of Java, and the OSes they offered (Linux, and SunOS/Solaris was eventually opened as well), and so on.

Re:Compensation for Java?

[poetmatt]

Note the updates from Groklaw. (link to article)

UPDATE

Just in case you missed it (or simply don't want to spend your time searching for it, there are at least three important takeaways conveyed in the Google brief and related documents:

        * Cockburn ignored prior negotiations between Sun and Google in which Google was offered the opportunity to license these and other patents for a fraction of Cockburn's present estimate;
        * Cockburn ignored other licensing transactions in which Sun licensed these patents for a fraction of Cockburn's present estimate (and these other licensing transactions will almost certainly become a limiting factor on any royalties Oracle may be awarded); and
        * Cockburn bases his estimate on worldwide sales of Android devices and Google revenue, despite the fact that the devices are made and used (and thus the infringement occurs) outside the U.S. and is not subject to a U.S. patent claim.

All of this serves to indicate that the Cockburn report, while sensational, has little or no bearing on a likely outcome of this case

Re:Compensation for Java?

[Anonymous Coward]

Hey maybe Oracle should be paying some of us damages for inflicting Java on the world. ;)

Or JDE...

Re:Compensation for Java?

[Anonymous Coward]

Or saving the world from the incomprehensible shit you produce that you 'call' C++

Re:Compensation for Java?

[Anonymous Coward]

Incomprehensible to you maybe.

Re:Compensation for Java?

[DrXym]

Hey maybe Oracle should be paying some of us damages for inflicting Java on the world. ;)

Java's not a bad language although the pace of development is glacial. No more so than on J2ME where Oracle / Sun did vastly more to damage the platform than any other company. J2ME was built for another age. It might still have application in PLC and lightbulb dimmers, but not in phones or set top boxes any more and I fail to see how Android "damaged" anything of theirs.

Companies were already deserting in droves before Android took off. If Oracle were smart they'd put all this shit behind them and try some constructive dialog with Google with the hope of getting something like JavaFX 2.0 and some other bits and bobs into the standard API stack.

Re:Compensation for Java?

[Anonymous Coward]

Still better than responsible-for-90%-of-the-word's-exploits-and-memory-leaks C.

Of course, compared to Haskell, every imperative language is a painful sorry piece of shit.

Re:Compensation for Java?

[Anonymous Coward]

I like the Carbon (C++) offsets.

jr

Re:Compensation for Java?

[TheRaven64]

A good programmer writes code than an average programmer can maintain.

Re:Compensation for Java?

[wurp]

either Oracle is right and thus is in violation of it's (Sun's) own terms for Java

That's not how licensing works. Licensing are the rules under which people who *are not the copyright holder* may distribute copies. If you own the copyright, you may distribute copies however you like.

I don't give up rights on what I can do with my copyrighted materials when I distribute under a license. I am just giving people the option to distribute in ways otherwise restricted by copyright law. I can't take away your distribution rights with a license, and I certainly don't give up my own distribution rights by granting you a license.

Re:Compensation for Java?

[ATMAvatar]

licensed per core x RAM - they don't care if it's a server or if you need a seat for a development or QA lab workstation they license it based on what a given CPU "could" theoretically handle.

Not that I disagree with most of your Oracle == evil rant, but they do have an Express Edition you can use for dev/QA work for free.

Re:Compensation for Java?

[Anonymous Coward]

Quit moving the average down.

Re:Compensation for Java?

[kimvette]

Not that I disagree with most of your Oracle == evil rant, but they do have an Express Edition [oracle.com] you can use for dev/QA work for free.

No, you can't. You really can't. You can develop on it sure but you can't really adequately rely on it for quality assurance purposes, other than compatibility.

Why, you ask?

How do you conduct a proper performance/stress test on that free edition?

First

[Anonymous Coward]

Don't develop with Java or you'll get fucking sued. Fuck oracle

Re:First

[Anonymous Coward]

More like don't develop a clone of Java targeted at the mobile platform, where Sun/Oracle made most of their Java licensing revenue, or you'll get fucking sued.

There is no issue developing in Java

Re:First

[nagnamer]

There is no issue developing in Java

Apart from Java itself, you mean?

New Google Strategy

[ackthpt]

1. Buy Oracle

2. Sack those who are responsible for the suit.

3. Open Java to the Public Domain

4. Sell Oracle.

Re:New Google Strategy

[Anonymous Coward]

Except oracle is the second largest software company in the world!. Pretty much impossible for Google to ingest oracle.

Re:New Google Strategy

[Anonymous Coward]

Google: Challenge Accepted!

Re:New Google Strategy

[wjousts]

Even if your strategy wasn't impractical, what would make you think that Google would want to make Java public domain?

Re:New Google Strategy

[Anonymous Coward]

Unfortunately GOOG ($156.17B) is worth less than ORCL ($160.98B) at the moment.
What you said can't happen now, but it could have been plausible at the end of 2007.

Re:New Google Strategy

[carlosap]

Oracle bought sun in 7.38B. Win 6.1 B for damages, priceless!, that was in oracle eyes before the adquisition. Quite a bargain to bought Sun!.

Re:New Google Strategy

[RockGrumbler]

http://royal.pingdom.com/2010/03/17/how-much-big-tech-companies-have-in-the-bank/

oracle is worth more then google on paper.

Re:New Google Strategy

[ackthpt]

Even if your strategy wasn't impractical, what would make you think that Google would want to make Java public domain?

Google, for all their recent goofs, still believes in making things available - grow through acceptance and use of technology, rather than standing over customers and developers with a fee schedule and a large club.

Re:New Google Strategy

[Amouth]

they only need to own 51%

Oracle's Market Cap = 161.13B

Price to control = 80.565B

Google's Cash on Hand = 36.67B

so yea they are short a bit..

Re:New Google Strategy

[Talderas]

Don't worry my selective reading lead me to read this line initially.

"In August last year Oracle sued Google for infringing Java patents and copyright by developing Android. Google argues that Android uses technology derived from Java and therefore infringes multiple patents."

I guess I was expecting Google's counter to the sue rather than more info from Oracle's perspective.

Also. Help me Slashdot. Which side am I supposed to hate? Google or Oracle?

Re:New Google Strategy

[v1]

they only need to own 51%

I think you meant to say "controlling interest"? If say, a company is owned equally by five groups, each with 20% of the stock, you could control the company by acquiring 25% of the stock.

Of course owning 51% guarantees you controlling interest, but strictly speaking, it's not necessary

Re:New Google Strategy

[Anonymous Coward]

Also. Help me Slashdot. Which side am I supposed to hate? Google or Oracle?

Oracle. I would have had to hesitate briefly if you'd said "SCO or Oracle".

Re:New Google Strategy

[i.r.id10t]

Which one didn't offer you a job or allow you to buy stock in the IPO?

Re:New Google Strategy

[BuckaBooBob]

It would be funnier if they Open-Sourced Oracle DB Kept Oracles Patent Portfolio and put all the different parts of Oracle up on EBay or Craislist.

Re:New Google Strategy

[140Mandak262Jamuna]

One small problem. Market cap of ORCL is 167 billion dollars. GOOG is 156 billion. Very curiously the difference seems to be twice the amount ORCL is dunning GOOG for. I leave it to the day traders to argue this is what the market is valuing the value of the law suite.

Re:New Google Strategy

[MicroRoller]

Doesn't make much sense to spend 10's or maybe over 100 billion dollars to acquire a company in lieu of being sued over 8 billion or trying to fight it or settle.

Better solution is if both Ellison and one of the Google founders agree to the take battle to the skies in their respective fighter jets to settle this. Winner also gets any adsense revenue from the YoutTube videos of the aerial combat.

Re:New Google Strategy

[wjousts]

Good guys in the business world are like the tooth fairy or Santa Claus. They don't fucking exist. Java already has wide acceptance, so why make it public domain?

Re:New Google Strategy

[obarthelemy]

Depends on whether the other shareholders agree with you ?

Re:New Google Strategy

[Anonymous Coward]

If say, a company is owned equally by five groups, each with 20% of the stock, you could control the company by acquiring 25% of the stock.

No you couldn't. Where do you get this crap from? 25% can be enough to control a company that has a diverse ownership who aren't likely to all get together and set a strategy, i.e. millions of small shareholders, but if there are only five groups holding it with 20% each then just two of those groups need to take enough interest to outvote your 25%. That's nothing like control.

Re:New Google Strategy

[kimvette]

Most of Java is licensed under the GPL to begin with - that's actually better than public domain in some ways, because this could be used to turn around and bite Oracle in the ass, since they are distributing GPL-licenses software, and the terms require it to not be hindered by patents else it can't be redistributed. So they could not only get hit by estoppel since they are in fact inviting people to infringe their patents, but those sued could conceivably turn around and countersue Oracle for immense damages that go beyond actual fiscal losses.

Re:New Google Strategy

[Anonymous Coward]

because java has wide acceptance, it should be made public domain.

Re:New Google Strategy

[wjousts]

That's not an argument. Why should it be made public domain? If I (as Google in this hypothetical situation) spend billions of dollars to buy Oracle, what do I get for making Java public domain? How does it add value to the shareholders?

Paying a large sum of money to acquire a company, dump one of it's prime assets for free and then selling it again (no doubt at a huge loss since it has less assets now) makes no sense at all. The shareholders would rightly crucify them.

Re:New Google Strategy

[TooMuchToDo]

You don't think Google could find enough other tech companies who want to cripple Oracle? Deep pockets + collective effort = collective ownership of controlling interest in Oracle. Google and Microsoft could team up, eat up Oracle, and Google could let Microsoft get all the database/middleware business for their troubles (with Google walking away unscathed).

Re:New Google Strategy

[Fantom42]

You forgot to sack the people responsible for sacking those responsible for the suit.

Re:New Google Strategy

[swillden]

Even if your strategy wasn't impractical, what would make you think that Google would want to make Java public domain?

I don't think Google would make the Java tools and libraries public domain. I think Google would release them under an MIT-style open source license, as they've done with the Go language.

Re:New Google Strategy

[Anonymous Coward]

Compare the prices from 10 years ago. Now where do you think Oracle with be in 10 or 15 years with generic Intel + Linux/BSD systems killing their market now, and the "cloud" looming in the distance? Oracle will need to get as much from patents, licensing and other IP as they can now, while they're still a massive company.

Re:New Google Strategy

[guybrush3pwood]

1. Buy Oracle

2. Sack those who are responsible for the suit.

3. Open Java to the Public Domain

4. Sell Oracle.

No need for #4. Once you bought Oracle, keep it and enjoy being the new Microsoft.

Re:New Google Strategy

[Anonymous Coward]

You neglected to mention that a takeover is likely to increase the stock price. It's not inconceivable that if rumors of a Google takeover came down the pipeline that the stock price would jump 10-20% simply based on rumors.

Re:New Google Strategy

[TheLandyman]

Off-topic.. I think Apple and Oracle should team up. I think there's writing on the wall, but I might just be crazy. 5 year Apple strategy for business share: 1) Partner with Oracle to use expensive Sun hardware and software in the data center 2) Sell expensive Apple hardware at the desktop. 3) ???? 4) Profit Given how much money IT departments waste these days (SOA anyone?) it should go over quite smoothly. Time to buy some Apple stock.

Re:New Google Strategy

[Anonymous Coward]

While your comment technically parses, please, for the love of $DEITY, learn the word "than".

Re:New Google Strategy

[mrwolf007]

So sure about that?

Who knows, Sun might have survived if Google had paid royalties.

Re:New Google Strategy

[GauteL]

Not sure. Doesn't the majority shareholder still have to act in the interest of the company? AFAIK there are laws as to what the majority share holder is allowed to do as to not ruin the investment that the minority shareholders have in the company.

If it is in Oracle's interest to sue Google for copyright infringement, then I'm not sure if buying 51% of Oracle would really allow Google to cancel that lawsuit.

Obviously this is all academic anyway.

Re:New Google Strategy

[TheRaven64]

For the same reason that making WebM BSD licensed after spending $300m to acquire it made sense: they can't make money from selling it, but they can make money from the surrounding ecosystem, and that ecosystem is enlarged if it's free.

Re:New Google Strategy

[ChunderDownunder]

android's use of the apache license aside, i'm sure Google would be happy enough with openjdk's GPL. The problem isn't the GPL but the additional conditions attached.

The real issues here are Sun patents, jck restrictions, royalties for 'embedded' Java and loading only a subset of Java SE on mobile devices.

Abolish the jcp, guarantee patent freedom, make openjdk a proper open source project - like red hat's icedtea. Modularize the jre - like Sun demoed in kernel mode on windows.

Re:New Google Strategy

[ChunderDownunder]

patent protection for GPL'd openjdk doesn't extend to cleanroom Apache Harmony - Gosling's/Larry's rules.

Re:New Google Strategy

[HappyEngineer]

If they were going to sell Oracle again they'd need to make it public domain (or something similar) to prevent the next buyer from taking control of it again. I suppose an alternative would be to buy Oracle, sell Java to Google for $1, then sell Oracle.

Re:New Google Strategy

[exomondo]

For the same reason that making WebM BSD licensed after spending $300m to acquire it made sense: they can't make money from selling it, but they can make money from the surrounding ecosystem, and that ecosystem is enlarged if it's free.

WebM was freely licensed because that was the only way it could compete with h.264, it had to offer something substantially advantageous, being 'just as good' isn't good enough to justify a switch otherwise they would just continue using h.264.

Re:New Google Strategy

[gmhowell]

Even if your strategy wasn't impractical, what would make you think that Google would want to make Java public domain?

Google, for all their recent goofs, still believes in making things available - grow through acceptance and use of technology, rather than standing over customers and developers with a fee schedule and a large club.

Yeah, like SketchUp Pro.

Ummm...

[MightyMartian]

Isn't this a standard tactic, ask for some obscenely high figure and then settle for much much less?

Re:Ummm...

[ackthpt]

Isn't this a standard tactic, ask for some obscenely high figure and then settle for much much less?

Know anything about Larry Ellison? He thinks he's a Samurai. You have underestimated his determination to bring Google to their knees.

Re:Ummm...

[Lorien_the_first_one]

Based on the arguments in the Daughbert motion posted on Groklaw, I think Oracle will have no choice. Google is taking a pretty good shot at Oracle's expert witness, and based on what I'm reading there, I think Google is likely to prevail in their challenge to the witness.

Re:Ummm...

[Amouth]

Right by Oracle is a blood money machine.. Google is a threat to them - if they have a chance at burning them i really down Larry is going to pass it down..

Also for how fast this happened after they bought Sun i wouldn't be surprised if this was planned long before they approached Sun with a buyout.

Re:Ummm...

[yog]

From the Groklaw article: "Cockburn Offers No Meaningful Analysis Regarding Copyright Damages"

That just about sums it up. Oracle shouldn't be picking a fight with Google; they should be thanking Google for helping to spread general Java know-how and promoting it on their phones, even if they've found a way to evade the licensing fees by using a 3rd party JVM.

Nokia has just started a partnership with Microsoft, so Windows Mobile and Bing Search will probably be their standard platform, with Visual C# as the primary language. Blackberry still uses Java, but they're going down the tubes as fast as Nokia. Meanwhile, Apple continues to prefer Objective C. That leaves only Android as the major handheld platform for any flavor of Java.

If Oracle wishes to spread Java on the handheld, they could maybe start by not suing the maker of Android. They should instead be negotiating with Google, trying to integrate Oracle services into Android, maybe offer Google a good deal on a fully licensed JVM that performs better than Dalvik. Wasting millions of dollars on lawyers and risking a huge schism with Google hardly seems worth it.

Microsoft is Google's rival; Microsoft is Oracle's rival. Increasingly, Apple is Google's rival. Maybe the two should get together and unite against Microsoft (and Apple, which has little invested in Oracle's product line). Stupid lawsuits, wasting everyone's time and money. How many programmers could they have hired for the amounts they're spending and will spend on this ridiculous effort?

Re:Ummm...

[JamesP]

Well

J2ME was a mess and a failure

Android 'fixed' Java on mobile, at what cost?!

Now it risks going down the drain again!

And Oracle doesn't care about mobile. Oracle wants the multi-million JVM corporate supports

Re:Ummm...

[Johnny Mnemonic]

They should instead be negotiating with Google
This is probably simply the first stages of exactly that. I fully expect a negotiated agreement in the end, that will mutually benefit both. And probably only wind up costing Google lawyer's fees, if anything. This is just Oracle reminding Google who was ultimate authority of Java, and to increase the negotiating position.

Re:Ummm...

[ColdWetDog]

Know anything about Larry Ellison? He thinks he's a Samurai. You have underestimated his determination to bring Google to their knees.

Eating sushi does not make you a Samurai.

Re:Ummm...

[bWareiWare.co.uk]

Dalvik doesn't have to fully implement the Java standard, so it can take shortcuts to be faster then any licensed JVM.

Re:Ummm...

[Anonymous Coward]

Are you kidding? This is oracle, the company headed by the man who is arguably the largest douchebag in the entire world. The man who hired HPs former CEO, who was outed for buying prostitutes with company money, just to spite HPs other board members.

Re:Ummm...

[im_thatoneguy]

This is like the tax debate.

Eventually that 'popularity' has to translate into revenue.

Sure the could give away Java and it would be hugely popular. But "Hugely Popular" doesn't make them any money. 0% of infinity is still $0.

Currently Google is paying $0 for each java license. If Java just becomes "The android language" and they continue collecting $0 why should they care.

It's kind of like the people who try to use photos without the photographers permission. "Hey but you're getting free exposure." "You know what's better than free exposure? Buying stock photos." Exposure is only worth something if it translates eventually into sales.

Re:Ummm...

[Anonymous Coward]

But eating... oh, yeah.

Re:Ummm...

[shugah]

But does eating sushi wearing a turtleneck make you a ninja turtle?

Re:Ummm...

[GooberToo]

Currently Google is paying $0 for each java license. If Java just becomes "The android language" and they continue collecting $0 why should they care.

People keep parroting this crap because Oracle keeps regurgitating this crap. It doesn't exist. Google doesn't need a Java license because they don't use Java. Developers use the java compiler to compile java byte code. That's it. Period. Google is using freely available libraries, which are written in the java language. Oracle's own compiler compiles it and that license is extended by Oracle to all developers. Google isn't a player in the least. At this point, Google's own tools convert the Java byte code into Dalvik byte code. Google's own tools and VM only ever deal with Dalvik, except for their one component which reads Java byte code (which is both documented and legally accessible under reverse engineering exceptions).

Basically, the fact people are so stupidly regurgitating Oracle's lies and starting the premise of defense on nothing but lies already says Google should be scared of Oracle - not because Oracle has any legitimacy, but simply because people are constantly proving two thing; one, they are really fucking stupid, and two, these really fucking stupid people are listening to Oracle.

Re:Ummm...

[ChunderDownunder]

Faster?? Benchmarks for dalvik vs hotspot for java se embedded showed hotspot was ahead on pretty much every score.
Dalvik's strength is in lower resource consumption and spawning new vms on limited hardware.

Re:Ummm...

[Alien Being]

Samurai? Is that another word for cunt?

Re:Ummm...

[im_thatoneguy]

My point wasn't whether or whether not Oracle has a case. The question is whether or not Oracle is *wrong* to attempt to extract licensing fees on principle.

On principle the idea that infinite free promotion is useful. I can't buy a lunch or make a car payment with self promotion.

Re:Ummm...

[shutdown -p now]

Oracle's patents are not for Java or JVM per se, though, but rather for some more generic VM implementation techniques that they claim to be applicable to Dalvik. It's hard to say whether they have a standing, but I've heard claims that those patents are broad and established enough that Microsoft licensed them from Sun (and now Oracle) to be able to implement .NET. And .NET is about as different from JVM as Dalvik is.

Re:Ummm...

[Anonymous Coward]

Microsoft is Google's rival; Microsoft is Oracle's rival. Increasingly, Apple is Google's rival. Maybe the two should get together and unite against Microsoft (and Apple, which has little invested in Oracle's product line).

And of late, HP is Oracle's rival (must remember to tick the anon box :)

Re:Ummm...

[nedwidek]

Dalvik doesn't implement any of the Java standard. They're not taking shortcuts to make it faster. Rather they went with a register based VM like Parrot, instead of a stack based VM like Java. The Dalvik bytecode files are bigger than the equivalent Java files as a result, but the fact that a register based approach is so much faster results in the Dalvik code still running faster.

Re:Ummm...

[GooberToo]

Whoosh. That's the point. A java license has nothing to do with anything. Their argument is about patents - not Java licenses. Period. You're a smart guy and even you are conflating the two. Furthermore, the patents which Oracle is laying claim to, are so broad and/or not even applicable, the patents should have never been granted and/or don't even apply to Dalvik.

The fact that someone as smart as you is conflating the two, completely validates my point. You're buying the bullshit from Oracle and don't even realize it.

Re:Ummm...

[shutdown -p now]

Their argument is about patents - not Java licenses. Period.

Again, I may well be wrong here, but isn't the whole point of getting the "Java license" (aside from being able to use the trademark "Java") is to be given a license to use those patents?

Or are you saying that Google doesn't need the "Java license" as such, but rather a different kind of license that would permit to use those patents in their (non-Java) products, like MS allegedly does for .NET? That's true, though I guess Oracle could just tell Google that the licensing terms for their patents is that the implementation must be certified as Java (i.e. they just can't get it for Dalvik, period) - they aren't really required to provide fair licensing terms...

Or are they deriving the amount of damages from the cost of their Java license, and you're saying that it's incorrect? I don't know how this is normally treated, actually - how do you calculate damages for patent infringement? "Fair market price" - how is that even computed? On the other hand, it clearly can't be whatever the patent owner says it should be...

Furthermore, the patents which Oracle is laying claim to, are so broad and/or not even applicable, the patents should have never been granted and/or don't even apply to Dalvik.

I sure hope so, but that is ultimately for the courts to decide. And, given how such things usually go in US these days, I'm not very optimistic about it, to be honest.

Re:Ummm...

[Anonymous Coward]

Someone mentioned in the comments of a previous article that this potential lawsuit was offered to potential buyers of Sun. It is more likely the initial planning was done by Sun before the buyout rather than Oracle.

Obligatory

[Spad]

Groklaw commentary

Re:Obligatory

[micheas]

The former head of RedHat's legal team took over groklaw.

Re:Obligatory

[Lanteran]

I'm thinking somebody else took up writing stuff for them after Pamela Jones quit.

Re:Obligatory

[ZombieBraintrust]

The creator handed it off to someone else so they could persue other interests. So it is basically under new management.

Re:Obligatory

[Anonymous Coward]

The former head of RedHat's legal team took over groklaw.

...took over from IBM's legal team?

This is what Google gets for using Java

[gatkinso]

C++ is looking pretty good right about now.

Re:This is what Google gets for using Java

[kvvbassboy]

C++ is and was always good. It's just that people can't program non-buggy in it, and to be honest it is relatively difficult to do so (which is what makes it fun if you have time ;) ). Try vala, although technically it compiles to C.

Re:This is what Google gets for using Java

[macintard]

Unless you want to develop on a platform that utilizes automatic garbage collection. You know, for security reasons?

Re:This is what Google gets for using Java

[Anonymous Coward]

No it is what they get for NOT using Java(tm). If they had of been using Java(tm) they would have gotten the patent grant that is associated with using Java(tm). Instead they used Harmony, which is not Java(tm).

Re:This is what Google gets for using Java

[Dog-Cow]

I hate C++ and Java both, but both of those platforms have garbage collectors available. It just so happens that only one has it by default with every practical implementation.

Re:This is what Google gets for using Java

[codepunk]

Yea because you know we have not seen any application security problems with any android app. Oh Wait.

Re:This is what Google gets for using Java

[softWare3ngineer]

It's just that people can't program non-buggy in it.

Its hard to write non-buggy code in any language. ill admit it is easier to mess up alot of stuff in c++, but bug free code is not often seen anywhere. A team must make bug free code a priority and devote alot of time to achieving that goal. Even then it might not be possible to write truly bug free code.

Re:This is what Google gets for using Java

[Anonymous Coward]

Wrong choice of words, garbage collection has no relevance to security, but basically helps keeps memory usage in check.

Re:This is what Google gets for using Java

[SplashMyBandit]

... and the fact the standard C++ libraries don't do much (compared to Java), and libraries that do stuff are generally no portable, and there is no in-language support for multi-threading, and ..... That is why C++ use declined and Java took over as King of the Hill in global development (according to the Tiobe Index at least).

Re:This is what Google gets for using Java

[Twinbee]

They would have been safer going with C# rather than Java. And it's better.

Re:This is what Google gets for using Java

[Anonymous Coward]

What a terrible argument. Garbage Collection != Security. At best it might save on RAM. It won't even close an open file handle for you. RAII is far more suited to the task.

Re:This is what Google gets for using Java

[Tarlus]

You know, for security reasons?

I know you ain't talkin' about Java!

Re:This is what Google gets for using Java

[Anonymous Coward]

memory leaks can be exploited to run arbitrary code so the whole memory management model, which includes garbage collection is relevant to security.

Re:This is what Google gets for using Java

[petermgreen]

Wrong choice of words, garbage collection has no relevance to security

Taking memory disposal out of the hands of the programmer makes it far more practical to prevent programmers accessing memory through stale pointers (either accidentally or maliviously). Preventing programmers accessing memory through stale pointers is important for any environment that mixes code of different trust levels (think the java plugin in your browser) and also helps limit the damage programmer errors can cause.

Reference counting could be used instead of tracing GC but in typical GUI code the programmer would have to use a lot of weak references to prevent dead cycles.

Re:This is what Google gets for using Java

[Anonymous Coward]

Except when you are talking about software that gets sold to end users, java is DOA and C++ is king.

Re:This is what Google gets for using Java

[Anonymous Coward]

You're thinking about pointer arithmetic and non-bounds-checked arrays. Not the same thing.

Re:This is what Google gets for using Java

[shutdown -p now]

If they had of been using Java(tm) they would have gotten the patent grant that is associated with using Java(tm).

You don't get a free patent grant for Java implementation running on a mobile device. It's only for the desktop.

Re:This is what Google gets for using Java

[JasterBobaMereel]

One Definitely patent encumbered language, rather than one they can probably work around the patents .. Hmmm ...

Re:This is what Google gets for using Java

[sjames]

Python would be even better for this. Just a little slower, but a lot more compact. But for maximum compatibility with everything, the underlying API should have been just plain C. Pretty much anything can interface painlessly with that, especially if you stay away from passing pointers in structs.

Re:This is what Google gets for using Java

[Twinbee]

As far as I know, I thought C# was a very open language, and that only Microsoft's implementation was protected. It would be easy for Google to write their own implementation.

So basically...

[itchythebear]

So basically Oracle thinks that Google is expecting to spend about 6.2 Billion dollars to pay it's lawyers and is hoping they see it would just be cheaper for them to license the technology from Oracle.

Please oh please let Google fight this and win. Not because I love Google, but because I hate Oracle.

Re:So basically...

[Anonymous Coward]

Yeah!
Oracle, it is just a rogue.

muahahaha

[Anonymous Coward]

oracle==evil

Re:muahahaha

[Anonymous Coward]

Unless you meant to test whether oracle is equal to evil, I believe you meant:

oracle = evil

Re:muahahaha

[The_Wilschon]

You don't need to set oracle to evil, it already has that value. You just need: assert oracle==evil

Good.

[Hazel Bergeron]

Oracle has had decades of experience at being a productive prick. Unlike Ballmer, Ellison is not an accountant upgraded beyond his station. If there's someone who will put up a determined fight to cut Google down to size and hinder Brin and Page's journeys into everyone's lives, it's him. If the option was available, they'd all be against the wall - but a second best is to have them squabbling with each other. Their egos are, I think, too great for them to engage in the usual cooperation-under-a-thin-veil-of-competition of large corporations.

(Soviet information-gathering was evil because that time round your family wasn't the one which got to protect its own privacy while taking everyone else's, right Sergei?)

Re:Good.

[Anonymous Coward]

Fuck Ellison. I hope he gets run over by a fucking bus.

Re:Good.

[Hazel Bergeron]

Ellison and B/P/S form two sides of the same coin, being corporations which have grown quickly and absurdly massive on the business of data gathering and mining. In the former case the link with government work was well-known; in the latter case it's met with cries of HURR CONSPIRACY THEORIST because we trust the government these days not to abuse any large repository of data and try really hard to ignore evidence.

But the individual people concerned are egomaniacal dicks, and that acts against their favour, which in turn acts in our favour.

Comment removed

[account_deleted]

Comment removed based on user account deletion

C++ always looked good.

[Anonymous Coward]

I have little sympathy for developers who just aren't talented enough to write maintainable code in C++ and then cry "C++ is at fault, not me, and interpreted languages are superior and the only correct way to do development because I can actually manage to write code that works in THOSE. Who cares about low-level performance optimization anyway? Having to free up memory yourself is too much to ask anyone."

Re:C++ always looked good.

[kvvbassboy]

You are probably a talented programmer, and I completely agree that C or C++ is the holy grail for low level optimization (if you are too noob for Assembly or Hex). I am personally fond of C++ and do not really "get" Java, although I do use it.

This may be obvious, but for many software it's a trade-off between development time, bugginess, manageability and speed, C++ is not always optimal.

Re:C++ always looked good.

[Anonymous Coward]

Oh, please. C++ has always looked like a train wreck.
May God have mercy on the soul of those who think that the painful process of learning to deal with it's numerous retarded idiosyncrasies in any way relates to skill or talent in programming.

Re:C++ always looked good.

[Anonymous Coward]

it's

As someone will inevitably niggle about this, I might as well get it over with: The possessive adjective is "its", without the apostrophe.

Re:C++ always looked good.

[zeroshade]

Every language has retarded idiosyncrasies, the skill and talent in programming is to learn a language and know how to properly use that tool to solve a particular problem. Well written C++ can be very elegant and not look at all like a train wreck. Poorly written code in any language is still poorly written code and will look like a train wreck.

Re:C++ always looked good.

[Duhavid]

What I consider funny is that the whole premise of how .net was sold to dev shops.
No memory management to have to deal with, right? Not so, at least for VB.net. Declare a form? Have to call dispose on it.
The third party grid control set we use, we have to call dispose on many things created for those grids.

Why "dispose" is ok and "delete" is not seems to have come down to "you bought the tool, suck it up".

Yeah, I get that dispose is there to clean up other allocated resources, but the memory is not freed either in those cases.
And RAII handles all cases without having to have dispose like kludges tacked on.

Re:C++ always looked good.

[geoffrobinson]

Kudos to you sir. You followed up on a fanboy debate over computer languages with a Grammar Nazi post.

Re:C++ always looked good.

[petermgreen]

No memory management to have to deal with, right? Not so, at least for VB.net. Declare a form? Have to call dispose on it.
The third party grid control set we use, we have to call dispose on many things created for those grids.

It's basically the same problem java has. Tracing GC is tolerable for memory management but is no use for managing other resources that likely need to be freed up sooner.

And RAII handles all cases without having to have dispose like kludges tacked on.

Afaict RAII is nice for the simple stuff but it does have problems of it's own. Primerally how multiple references to the same resource and reference cycles are handled. Get this wrong and you can end up with stuff not being disposed of because a reference cycle keeps reference counts nonzero even after all external references are gone.

Bad publicity

[danbuter]

If I was in charge of a software company, I'd think very, very hard about using something other than Java for any future projects. This lawsuit will quite possibly bite Oracle in the ass over the long term.

Re:Bad publicity

[Anonymous Coward]

Having almost completed a 4 year project using Java, I'm determined to never, ever make that mistake again. It's just plain awful, rediculously tedious and verbose beyond comprehension. I hope it dies sooner than Cobol before it did.

Re:Bad publicity

[Anonymous Coward]

They seem to be trying to alienate any sort of partners they have. That will sink them faster...

Re:Bad publicity

[Anonymous Coward]

i'm using it since 1996 and never thought that way, one of the wonderful language ever written. I serious doubt you even used it!!!

Re:Bad publicity

[homey of my owney]

Ummm, Cobol's dead? Cobol 2002 Standard

Re:Bad publicity

[Anonymous Coward]

Just because it took you 4 years to complete that "Hello World" application in Java is a reflection of your skills more than a reflection on Java :) Oh, and by the way, Cobol is is still around and going pretty strong after over 50 years, "sooner than Cobol" would be quite some time.

Re:Bad publicity

[shutdown -p now]

What other languages have you used to conclude that it's "one of the most wonderful languages"? I mean, you have to do some comparisons to make such a conclusion.

Or was it COBOL before?

Somebody needs a bigger boat

[Bloodwine77]

I guess Larry Ellison is needing some extra money to start building his next yacht.

Where's PJ when we need her?

[jenningsthecat]

It may be time for a Groklaw comeback...

Re:Where's PJ when we need her?

[Liquor]

PJ is still backing the site, and keeping an eye on things, but Mark Webbink (from Red Hat's general counsel) is doing the the articles.. and seems to be as impartial as PJ, though not with the same humour (so far, at least).

The articles there go into details about Oracles claims and Googles counters.

Re:Where's PJ when we need her?

[Anonymous Coward]

http://www.groklaw.net/article.php?story=20110619135516162

Re:Where's PJ when we need her?

[Anonymous Coward]

Mark is doing a good job. He's not PJ but his analysis is discussion addresses the critical aspects that his reading audience needs to hear.

That said, I miss PJ too. She is Groklaw to me.

Oracle is going to get cock-burned by Cockburn. He is outright MAKING THINGS UP just so Oracle can throw around billion-dollar figures to the retarded technical press organizations.

I doubt...

[mr_lizard13]

I doubt Google would infringe on someone else's patents. That would be evil, and that's not how Google rolls.

Re:I doubt...

[Anonymous Coward]

ROTFL.... guys, dis is business not wonder land.

Google started abusing Java before the Sun acquisition, when Sun was in no position to fight them. Years before than, MS did in fact license some java related stuff to be able to release .Net.

IMHO Google is the evil one here and they should pay.

Of course it's possible that these numbers are high because settling the case, properly licensing Java ME for Android and bring Android back under the Java umbrella might be the preferred outcome.

In any case, don't forget that Oracle is the company who sued and got $1 billion from SAP for a much lesser infringement!

Re:I doubt...

[Anonymous Coward]

and how exactly that would be evil?

Fuck Google(.)

[Anonymous Coward]

Go Oracle!

Python for Android ... FTW!

[Skapare]

And I hear that Google has a lot of Python running in-house already. But if fewer CPU cycle per function performed is the goal for low power mobile devices, why not just plain old C?

Re:Python for Android ... FTW!

[Liquor]

Simple - Google needed app developers to write for the platform, and phone app developers mostly speak Java. So Google came up with the Dalvik VM that can run recompiled Java code in a compatible environment.

Re:Python for Android ... FTW!

[Issarlk]

Because if you require developpers to write in plain old C, there'll be a magnitude less apps available for Android.

Re:Python for Android ... FTW!

[mattdm]

How about requiring them to write in Objective C?

Re:Python for Android ... FTW!

[codepunk]

Oh, you mean like the lack of apps for iOS. Oh wait.

Re:Python for Android ... FTW!

[Nimatek]

Or why not assembly? The answer is, because developers need to get things done.

Re:Python for Android ... FTW!

[spintriae]

I wouldn't be surprised actually if Google was developing Go for the explicit purpose of replacing Java on Android. They announced Go a few months after Oracle acquired Sun, and last month at Google I/O, Rob Pike had this to say about it.

Re:Python for Android ... FTW!

[Anonymous Coward]

Python? Seriously? V8 (Google's JavaScript engine) is a lot faster and there are also more people who already know JS.

Re:Python for Android ... FTW!

[spintriae]

And perhaps most telling of all is that Go has supported ARM processors since day one.

Re:Python for Android ... FTW!

[wumpus188]

Looks like you're either mistaken or lived under a rock. Most mobile apps devs speak Objective-C now and it has been this way for quite some time.

Re:Python for Android ... FTW!

[Anonymous Coward]

The whole reason google used a vm in the first place was so that one bytecode could run on any type of supported mobile platform ARM VIA Intel. If a developer programmed in a language that complied to machine code it would have to be recompiled for every platform. The davak VM introduces simplicity by handling different hardware for the programmer and ease of development with the java language. Even the NDK (native development kit) which allows for c and C++ code to be used in an android app still complies into davik bytecode so that the program can be ran on any supported arch. Google uses the java language for android but not the jvm which does not violate oracle copyright or patients.

Re:Python for Android ... FTW!

[JonJ]

Yeah, about that... People wrote mobile phone apps years before iPhone. Fanboy much?

Re:Python for Android ... FTW!

[jepaton]

The limitations of compiled executables are becoming more apparent with today's diverse hardware. One generation of mobile device may not use the same instruction architecture (processor type) as the next generation of mobile device. And the other devices connected to the processor change frequently. A new executable would have to be compiled for every major variant of the device. With something like Java bytecode the program can be one-time optimised when the program is loaded onto the device, which is a good comprise between efficiency and portability.

I program in C. In C there is normally a good correspondence between C code and machine code. But there are many ways that C is less efficient compared to more modern languages, especially when writing well structured code.

Re:Python for Android ... FTW!

[Anonymous Coward]

if fewer CPU cycle per function performed is the goal for low power mobile devices, why not just plain old C?

Why not plain old ARM assembly? Java + Linux (with Eclipse as the IDE) is a fantastic combination... none of the technologies is perfect but it just works. I'm certain that combination was critical to the success of Android. Unfortunately the Linux "desktop" community never bothered trying something like this... and so far it's been a failure (compared to Linux in other segments).

Re:Python for Android ... FTW!

[hazydave]

Google wanted the basic Android apps to be machine independent. So the same stuff largely runs on ARM and x86... even a few MIPS fans are starting to do things with Android. Java was available, and of course, the language itself couldn't be patented or copyrighted, and there was this GPLed implementation, etc. Sure looked like a good idea at the time.

I had the same kind of plan when developing a video/internet set-top box back in the late 1990s. We wanted applications, but they needed to be insulated from the hardware, so in the future, the hardware could change completely without messing with compiled apps.

Sure, they could have written their own language as well as their own VM... would that have been enough to keep out of the crosshairs of companies like Oracle who now think they have ownership over any software VM, thanks to Java? Maybe, maybe not.

Re:Python for Android ... FTW!

[hazydave]

Looks like you're either mistaken or lived under a rock. Most mobile apps devs speak Objective-C now and it has been this way for quite some time.

True only for very small values of "quite some time" and "most mobile apps devs". Apple iPhone apps have only been "native" in Objective-C for three years and three months... the SDK was released in March 2008.

I guess YOU lived under a rock prior to then. Blackberry ran Java MDIP as far back as 2003, their basis for mobile apps. Countless "not so smart" phones ran Java ME. Nokia, Microsoft, and PalmOS smartphones and PDAs, dating back to the early 1990s, were usually programmed in C or C++, though plenty of other options existed. And even today, you can program iPhone apps otherwise... for example, in Adobe ActionScript or Flex. Check this out: http://en.wikipedia.org/wiki/Mobile_application_development

There is and has been a ton of mobile development unrelated to Apple's 10% or so share of the global mobile market (factoring in smartphones, iPads, and iPods).

Re:Python for Android ... FTW!

[benhattman]

Other posters already said it, but why not Go instead of Java? Instead of a virtual machine, just distribute the source and let a compiler optimized for that specific phone build it into an executable when it's loaded? One of the main points of Go is how fast it compiles.

I hope...

[noTimeAtAll]

I hope LulzSec makes some funny stuff to Oracle's servers soon.

I want some of those drugs

[msobkow]

I dunno what Oracle's people are smoking, popping, or injecting, but I want some!

Re:I want some of those drugs

[ColdWetDog]

I dunno what Oracle's people are smoking, popping, or injecting, but I want some!

From the squirrel-crazy behavior they've been displaying, it appears to be methamphetamine. That's bad stuff and you want to stay away from it.

I mean, just look at Ellison? Does he seem OK to you?

Re:I want some of those drugs

[Lifyre]

Those must be dentures then!

Re:I want some of those drugs

[Anonymous Coward]

Ellison looks like someone I wouldn't let near my kids!

uncle larry

[Anonymous Coward]

I have to say I was upset to see Uncle Larry as the worlds "second" highest paid CEO. Uncle Larry, we want to see you #1. You are the fucking man!

To

[ajzimm3rman]

make up for all our lost open sourcers of course!

They should not have used Java

[Un pobre guey]

Basing Android development on Java was a mistake. It is GNU/Linux, after all, and people should have used what they pleased to develop under Xfce, Meego, Gnome, KDE, or some other relatively well established GUI. Writing apps with C++/Qt would have been easier than Java, with faster executables and smaller memory footprint.

Re:They should not have used Java

[Em Adespoton]

For that matter, if they want to use a runtime library, why not go with WxPython, or even a DroidPython? You get the runtime protection of Java, the OO environment, the fast engines, small footprint, and avoid licensing issues.

I guess Google figured there was already a large body of commercial Java developers that would jump at the chance to develop for Android.

Re:They should not have used Java

[SplashMyBandit]

You jest right? Java has a huge number of libraries, tools, and developers available. This is why Google chose it. Java is also GPL, thanks to Sun. This is a battle over ridiculous US patent laws, not over copyright conditions.

Re:They should not have used Java

[Anonymous Coward]

and people should have used what they pleased to develop under Xfce, Meego, Gnome, KDE, or some other relatively well established GUI.

Yes, and then Android on your phone could have been as successful as Linux on the desktop!

Re:They should not have used Java

[Anonymous Coward]

Google's NIH is biting them in the ass now...

Re:They should not have used Java

[Un pobre guey]

I know, I know. Have you had a look at Qt? It compares quite favorably.

Re:They should not have used Java

[Twinbee]

Yeah, it's the one thing that's putting me off Android. If they'd have gone for C# on the other hand....

Re:They should not have used Java

[Un pobre guey]

It would have been the Year of the Linux Desktop! Um, on a phone.

Re:They should not have used Java

[Anonymous Coward]

Basing Android development on Java was a mistake. It is GNU/Linux, after all, and people should have used what they pleased to develop under Xfce, Meego, Gnome, KDE, or some other relatively well established GUI. Writing apps with C++/Qt would have been easier than Java, with faster executables and smaller memory footprint.

Linux is a kernel. Android puts a very different userland on top of that kernel, because choices that make sense for a workstation or desktop computer do not make sense for a cell phone. Which revisions of what instruction set should they have supported native compilation to? Do you expect app developers to recompile for different phones?

Re:They should not have used Java

[Un pobre guey]

Do you expect app developers to recompile for different phones?

Yes, a very real and practical part of using Qt libraries. It would be a minor obstacle.

Re:They should not have used Java

[Anonymous Coward]

The very definition of Java is write once run anywhere. That's WHY higher-level languages have been created.
Java is not the only one, it was just the most structured, strict, along with OOP.
The idea is to write the underlying OS (Linux), have runtime environment, and run the rest on that.

Re:They should not have used Java

[m50d]

Even if all that were true (and it isn't), needing to recompile your programs separately for each processor out there would kill such a platform dead.

Re:They should not have used Java

[salesgeek]

Google's idea with using Dalvik (and the Java language) was compile once, run on many devices. For the most part, if you stick to specs and not get to crazy with GUI and hardware, it kinda works... not as well as Google had planned... but it has made it easier for developers to target different many dissimilar phones without having to compile for each device...

Re:They should not have used Java

[SplashMyBandit]

The results of Qt are nice but the development model is truly awful. It combines the macro madness of C with the meta-compiler madness of IDL/Corba and all the semantic shenanigans of C++. Sorry, I'll take fully hardware-shader accelerated Java Swing any day, and I'll have it working nicely in my multi-threaded apps too, thanks. Remember that "less is more", by which I mean the relative simplicity of Java is a real advantage in the real world (although as a result Java is 'dissed' due to the snobbery of many code monkeys out there - as if adding keywords and constructs ad-naseum made anything better or more maintainable).

Re:They should not have used Java

[Anonymous Coward]

Yeah because we see how big success Linux on desktop has been.. especially the Linux app UIs. Nobody actually cares that Android runs on top of Linux.

By choosing Java Google got million of Java devs on board without having to learn new language. It's the smartest thing they did.

Re:They should not have used Java

[shutdown -p now]

As a user, I don't really care if your Swing app is hardware-accelerated or not, because it still looks like crap (esp. the file open/save dialogs; or ever tried resizing a window?).

SWT is okay, though.

Re:They should not have used Java

[SplashMyBandit]

Oh, so you're very out of date, and have never heard of or used a "Filthy Rich Client" then, and only used Swing apps written by monkeys a decade ago. That's a shame. A Swing app (or any app really) written by a skilled dev skill kicks the web's ass for usability and flexibility (just not for deployment).

Re:They should not have used Java

[shutdown -p now]

The last Swing app that I used was Netbeans. This was a couple of days ago.

I repeat: Swing still sucks because it doesn't feel native. On Windows, there are no proper top menu animations. Resizing windows leaves black holes near the edge which blink out as they are redrawn once layout is reflown to catch up - it's brief, but highly visible. Text boxes don't have proper context menus. And open & save file dialogs are plain abominations - there is a reason why the OS provides stock ones, and third-party extensibility is a big part of it.

kicks the web's ass

Where did web come into this discussion from? I wasn't comparing Java Swing apps to web, I was comparing them to Java SWT apps and Qt apps.

Did you, by chance, confuse SWT with GWT? Or do you not know what SWT is?

Re:They should not have used Java

[Anonymous Coward]

This is a rare opportunity for some reverse pedantry. Android is just Linux not GNU/Linux - it uses the linux kernel but not the other elements of the GNU system. Here's the relevant question from Stallman's GNU/Linux FAQ.
http://www.gnu.org/gnu/gnu-linux-faq.html#justlinux

Re:They should not have used Java

[SplashMyBandit]

Mate, on one application we developed I had to develop a couple of SWT custom controls and fix a few buggy third party ones. SWT is a mess on the inside and far harder to extend than Swing controls. I guess if you are on Windows only (who is these days?) then SWT might be ok for you. In our company's experience SWT is awful and we'd rather not use it, thank you very much.

I find it bizarre you think Swing sucks just because Netbeans is slow to draw (although in versions since 6.0 I never notice on my Mac, Linux or Windows machines). I was on the Netbeans open beta team for 7.0 and suprised you have this slowness in drawing, I can't recall anyone else having the same issue. I wonder if there is something dodgy with your particular system (software rendering?) since *all* of Java 2D is hardware rendered by DirectX or OpenGL shaders (depending on platform). You simply can't get faster drawing than that (unless you somehow have a Sun JRE/JDK before 1.6.0_u10 lying around on your path somewhere) - although it is still possible to write sucky code if you don't know what you are doing.

Re:They should not have used Java

[shutdown -p now]

I guess if you are on Windows only (who is these days?)

About 80% of desktop users?

(note: I didn't say that I'm only on Windows - I also have a Mac laptop - but it is my primary desktop OS, so something that sucks there is not an option)

I find it bizarre you think Swing sucks just because Netbeans is slow to draw

You keep missing my point. Nowhere in my post I said that it's slow to draw. What I said is that it doesn't look and feel native.

One particular case where it is "slow" is when it comes to repainting windows as they are resized. A simple test case: open Netbeans, and drag any corner of the window quickly. You'll see black stripes at the edges of the window being moved. This happens not because it is slow to draw, but because Swing layout lags behind OS resizing the window - once you resize, there is a short (but noticeable) period of time where the contents itself is still laid out with the old size, and painted as such, including the root widget which is supposed to fill the window client area - and thus you get those unpainted areas which are just filled black. This has zilch to do with hardware acceleration of rendering. And any Swing app does that.

Some examples of where it is just not native:

1) Open Netbeans and mouse over any menu item in the top-level menu. Watch the dropdown open. Notice the lack of shadow (compare to Notepad to see what it should look like).

2) Right-click in any textbox or combobox control. Notice the lack of any context menu. Now press Win+R and right-click in the combobox in the dialog window that opens - that's what it should look like.

3) Find some non-editable combobox (the one that looks like a button with an arrow on the right in Windows). Look at the corners. Compare to the same widget in any native app.

3) Go to File -> Open File and look at the File Open dialog. Now do the same in Notepad, and compare the dialogs side-by-side. It's already pretty glaring (esp. on Vista/7), but also try right-clicking on the files in file selector and look at the differences in context menus that appear.

etc. Note that all of these are also reproducible in SwingSet2 demo with Windows theme enabled, so this isn't really about Netbeans.

For the record, this all is on JRE 1.6.0_24 on Windows 7.

Re:They should not have used Java

[SplashMyBandit]

> About 80% of desktop users?

This statement is true but myopic. Why only target 80% of a slice of the market. Why not choose a technology that you can use to develop for 80% of the market (including web, desktop, enterprise, phone) where only the front-end toolkit changes and the business logic remains the same. Unfortunately you seem focussed on optimizing your tactical technical objectives rather than strategic business ones (if I only considered the tactical I would agree with you). Also, looking to the past (the desktop) means you'll miss out on using your current code investment in the future growth areas. That's why I choose Java (plus Swing and GWT, and is easy enough to move to Android).

The 'its not native' is also another fallacy since there is no such thing as a 'typical Windows program'. Users actually don't give a frig. They care if your software is slow, buggy, has less features than another program, or was written by a developer who is clueless about good UI design and what users need to do. Ugly software has been commercially successful and kinda useful: look at all those hardware-manufacturer supplied interfaces that are basically controls on a non-resizable bitmap. How many users whinge about those interfaces not being native? What about the fact that every version of MS Office looks and feels different, do people whinge about that and not buy it as a result. Clearly you don't like some Java program you have used and try to rationalize your dislike, as if it was fundamental to the technology. Unfortunately, desktop-focussed folks like yourself don't ever seem to be aware that in the Enterprise there is a huge amount of Java software (on the desktop, intranet and back-end). No surprise, since companies don't advertise their internal software (for the effort, and because of possible competitive advantage). Again, if it was a Windows-only world I might agree with your position, but the reality is both Windows is not going away but also it is becoming less and less relevant. Targeting a single platform with your codebase, when you can just as easily target many, makes no business sense (and actually, never did).

Re:They should not have used Java

[shutdown -p now]

Look, I'm not complaining about Java here. I'm complaining about Swing. I don't care how it looks to a developer, but as a user I refuse to use Swing apps unless nothing else is available - and usually, there is.

And yes, there is such a thing as a "typical Windows program". It's an evolving definition - e.g. Ribbon wasn't part of it 5 years ago but is now - but it exists.

True, as a user I care more whether your software is slow (or looks like it's slow) than whether it's pixel-perfect in matching the OS theme. But Swing programs do look slow. I already gave you one example - resizing the window. There's no excuse for that kind of graphical artifacts when performing a common UI operation in 2011. None. And yes, it is very annoying.

Some other features which I've listed have more to do with availability of features than pure look. For example, native file open dialog lets you do much more than Swing one, because native dialog has all the shell extensions enabled in it. It's practically a mini-file manager. And that can be very convenient at times.

As for popularity... so long as we're exchanging anecdotes, here's mine: I have seen exactly one internal app written in Java+Swing in several companies that I have worked to date. Most just use web apps. Most internal desktop apps that I've seen so far were written in .NET. For intranet web apps, it's about 50/50. Note, this is "enterprise", not "desktop-focused" world that you assume I represent.

Re:They should not have used Java

[SplashMyBandit]

Try something like Visual Paradigm. Beats the snot out of Windows-native crap like Enterprise Architect. And no, most Windows apps still don't use the ribbon and not every one of them is written in C#.NET (which I presume you mean by .NET). There is a huge amount of legacy VB stuff out there. And yes, most stuff is being written as web apps these days (if they don't need much functionality), but that is even less of a reason to choose Windows-only technologies, isn't it?

Re:They should not have used Java

[shutdown -p now]

And no, most Windows apps still don't use the ribbon

True, but changing fast now that Ribbon is available out of the box in all development frameworks shipped by Microsoft (MFC for native, WPF/Silverlight for managed), and that other MS-made apps all get ribbonized - even something as trivial as MSPaint.

not every one of them is written in C#.NET (which I presume you mean by .NET).

There's no such thing as "C#.NET". The platform is called .NET, the language is called C#, and Microsoft's implementation of the language is called Visual C#. There were products called "Visual C# .NET 2002" and "Visual C# .NET 2003", years ago, but I don't think you were referring to them.

The reason why I wrote ".NET" and not "C#" is because it's not worth the point to make any further distinction. An application written for .NET in VB is not really any different from an application written for .NET in C# - they compile to (mostly) the same bytecode, and use the same libraries.

There is a huge amount of legacy VB stuff out there.

There's a significant amount of VB6 code (which is why VB6 runtime is still officially supported as of Win7), but it's not anywhere near as big as it was 5 years ago. Most VB6 projects were migrated to VB on .NET ages ago, and that, as noted above, is really just a different syntax for C#, for the most part.

And yes, most stuff is being written as web apps these days (if they don't need much functionality), but that is even less of a reason to choose Windows-only technologies, isn't it?

It depends. When your intranet is already all MS (IIS, MSSQL, SharePoint etc) - which is true for a lot of companies out there - you don't really lose anything, since you're already locked in; and you get easy integration with all that stuff, and good development tools. Whereas if your corpnet runs Oracle (or Postgres) and Apache and whatnot, then yeah, you should probably consider Java first.

Either way, I don't have anything against Java on servers - it has performed admirably there for years, and laid groundwork for most web development methodologies in use today. If given the choice, as a developer, I'd still prefer C# on something like ASP.NET MVC, but that is purely a language preference - I want my lambdas, type inference, and syntax comprehensions, and Java feels positively antiquated after writing C# for several years. They're catching up, but slowly (lambdas are only coming up in Java 8). But then of course there's also Scala, which can be used with existing Java frameworks...

Re:They should not have used Java

[SplashMyBandit]

> True, but changing fast now that Ribbon is available out of the box in all development frameworks shipped by Microsoft (MFC for native, WPF/Silverlight for managed), and that other MS-made apps all get ribbonized - even something as trivial as MSPaint.

Ah, I see you are talking about new stuff. I'm talking about the *huge mass* of applications that users are using today.

Nope. C# as implemented in Mono does not look like C#.NET, hence I use a qualifier.
> It depends. When your intranet is already all MS (IIS, MSSQL, SharePoint etc) - which is true for a lot of companies out there

That's the small and medium enterprises youare talking about. Big, established companies (government departments, defences, banks) have a lot of everything - which no one person gets to choose anymore. Hence, developing Windows-only is a non-starter at the higher end of the spectrum.

> Either way, I don't have anything against Java on servers - it has performed admirably there for years, and laid groundwork for most web development methodologies in use today. If given the choice, as a developer, I'd still prefer C# on something like ASP.NET MVC, but that is purely a language preference - I want my lambdas, type inference, and syntax comprehensions, and Java feels positively antiquated after writing C# for several years. They're catching up, but slowly (lambdas are only coming up in Java 8). But then of course there's also Scala, which can be used with existing Java frameworks...

Ah, the language lawyer person emerges - must have the latest shiny eh? For big to huge teams on long-running projects these features are a business liability (big teams *must* plan for turnover and recruitment of less than A+ devs, except in very exceptional circumstances), plus your MS-provided tools must change every year or two or MS can't demand revenue for them. Again, your arguments are focussed on the tactical and not on the strategic.

Re:They should not have used Java

[shutdown -p now]

Nope. C# as implemented in Mono does not look like C#.NET, hence I use a qualifier.

Again: C# is a language. It is implemented the same in Mono as it is in Visual C#. The differences are in available libraries, which are part of .NET - which is not C# specific.

If you would like to dispute this assertion, then I ask you to provide examples of specific differences in C# language constructs as implemented by Mono and VC#.

Even then, if you want to refer specifically to Microsoft's implementation of C#, for whatever reason, it's called "Visual C#".

That's the small and medium enterprises youare talking about. Big, established companies (government departments, defences, banks) have a lot of everything - which no one person gets to choose anymore. Hence, developing Windows-only is a non-starter at the higher end of the spectrum.

Funny you should say that. Of five banks that I've used in my life, two ran their Internet banking on MS tech - one on ASP, another directly on ISAPI.

Ah, the language lawyer person emerges - must have the latest shiny eh? For big to huge teams on long-running projects these features are a business liability (big teams *must* plan for turnover and recruitment of less than A+ devs, except in very exceptional circumstances),

It's silly to consider lambdas and other such features as "new and shiny" in 2011. Even in C#, they have been in the language for 6 years now - and it's by far not the first mainstream language to have them, as that would be JavaScript (and maybe even Smalltalk - that'd push it back even further). Back when Java 1.0 was released, yeah, it was still a "new thing", and the argument along the lines of "we're not gonna find A+ devs to handle that kind of advanced stuff" held water. Today, the only mainstream language that still doesn't have lambdas is Java, and even that is scrambling to get them.

The only people who are pushing back against it are the same kind of people who were pushing back against OOP in 90s because "conditionals and loops are good enough for everyone, and we need to keep it simple", and before that other people like that were pushing back against structured programming because "gotos are good enough for everyone, and we need to keep it simple". Such people, when they refuse to move on with the times at all (rather than merely being conservative, but still moving ahead), are generally called "luddites". Luckily for us, the majority of people in the industry aren't luddites, which is why we have real and significant advancement in both programming languages and frameworks.

plus your MS-provided tools must change every year or two or MS can't demand revenue for them.

They must? Funny, that - I wrote a WinForms app for a company once, back in 2002, when .NET was still all new and shiny, having been released that very year. It still works fine, after 2 OS upgrades (98->XP->Vista), and I still occasionally make fixes for them - without porting anything to new frameworks and such. It compiles just as fine in VS2002 (the very first one) as it does on VS2010.

Again, your arguments are focussed on the tactical and not on the strategic.

My arguments are focused on pragmatic. It doesn't make sense to avoid lock-in if you do so at the expense of higher development cost. It may well be cheaper to rewrite the whole thing from scratch in mature tech of the day once every ten years, rather than hire gray-bearded programmers for $250k who are the only ones who still remember how, say, AWT works, to support an aging codebase that's hack upon hack upon hack (as these things tend to happen in any codebase over the years, no matter how well designed initially).

Re:They should not have used Java

[SplashMyBandit]

>They must? Funny, that - I wrote a WinForms app for a company once, back in 2002, when .NET was still all new and shiny, having been released that very year. It still works fine, after 2 OS upgrades (98->XP->Vista), and I still occasionally make fixes for them - without porting anything to new frameworks and such. It compiles just as fine in VS2002 (the very first one) as it does on VS2010.

Your app can't have been doing very much. Just routine form stuff I guess. Yeah, I also wrote WinForms apps starting back around the same time you did. It was ok for simple stuff, but a bitch to extend (just like your beloved SWT). Turns out that the Java Swing apps I was also writing at the time (although they were dog slow back then, thanks to the slow JVMs of that time) still also work - on all platforms too.

> Such people, when they refuse to move on with the times at all (rather than merely being conservative, but still moving ahead), are generally called "luddites". Luckily for us, the majority of people in the industry aren't luddites, which is why we have real and significant advancement in both programming languages and frameworks.

Nope. Such people use it because it still works (like C!) and the added complexity doesn't add much, especially like when in C# they add and deprecate features based on the fashion du-jour. Incidentally, you are living in a bubble. The numbers speak against you, please check out http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html for a more realistic appraisal of what is going on globally.

Well, I don't have the grey beard yet but I do have the salary (and the PhD in Astrophysics, incidentally). I realise that good design seeks to remove complexity, not introduce it arbitrarily because it piques the interest of an otherwise bored developer eg. those that change languages like they change their underwear in the misguided belief that a handle of extra language constructs will make up for a breath of solution-space libraries - which is where real gains productivity are achieved. Unfortunately many people don't understand great design (in any area) is generally about removing unnecessary complexity whenever they see it. As Einstein is often (mis-)quoted, "As simple as possible, but no simpler". Java is like C, it kicks ass because the language just enough to get the job done (and the real work is done by the libraries).

Incidentally, if you don't think platform independence matters from a business perspective then you won't even see the opportunities you've missed (I guess mere employees don't even looking at this aspect). Check out this interview where Austin Meyer of X-Plane fame personally made $3.5 million in one month because he chose OpenGL over DirectX (I already know what technology you'd advocate, but then you'd miss the $3,5 million personally in your pocket if you did): http://techhaze.com/2010/03/interview-with-x-plane-creator-austin-meyer/

Re:They should not have used Java

[shutdown -p now]

Your app can't have been doing very much. Just routine form stuff I guess. Yeah, I also wrote WinForms apps starting back around the same time you did. It was ok for simple stuff, but a bitch to extend (just like your beloved SWT). Turns out that the Java Swing apps I was also writing at the time (although they were dog slow back then, thanks to the slow JVMs of that time) still also work - on all platforms too.

Good, so we can agree then that, as far as getting old stuff working, WinForms is not any worse than Swing.

By the way, your attempts to refute my examples seem to suffer from the "no true Scotsman" fallacy: if a company has intranet built on Microsoft products, then you say that it cannot be a big company, and when an application is written in WinForms you claim that it's not a big app. No surprise then that "big" stuff only runs Java in your world!

Also, I cannot claim any love (or hate) for SWT as a developer simply because I haven't written any app in it. All of my good attitude towards SWT and corresponding bad attitude towards Swing come from my experience as a user of apps written in either. That said, I'll take a Qt app over both any day.

From developer perspective, I won't dispute that Swing as an API is superior to WinForms. It's actually designed around model/view separation, which is a good thing. But it's also antiquated now that all new UI frameworks not only do the same, but also promote declarative definition of view, and strict separation of markup (for views) and code (for models). See WPF, Silverlight or (if you insist on portability) Qt Quick for a good example of a modern UI framework.

Nope. Such people use it because it still works (like C!) and the added complexity doesn't add much, especially like when in C# they add and deprecate features based on the fashion du-jour. Incidentally, you are living in a bubble. The numbers speak against you, please check out for a more realistic appraisal of what is going on globally.

TIOBE is actually very flawed as far as ratings go, since it indexes search engine counts and not (e.g.) jobs or other meaningful metric. It's also specifically flawed if you're trying to do a Java vs .NET comparison that way, since, for starters, you'd need to account both C# and VB in .NET camp, and then you need to account standalone mentions of ".NET" without further clarification of the language, which is exceedingly common. In case of Java, this is all accounted under "Java", since both platform and language share the same name.

Anyway, if you look at jobs instead (counts and $$$), the picture is very different. There, it is largely region-specific - e.g. Europe tends heavily towards Java, Australia & NZ towards .NET, and North America is the battlefield. If you drill down deeper, variations exist even between US states. But, all in all, Java and .NET are roughly equally popular in enterprise today, and pay equally well.

I realise that good design seeks to remove complexity, not introduce it arbitrarily because it piques the interest of an otherwise bored developer eg. those that change languages like they change their underwear in the misguided belief that a handle of extra language constructs will make up for a breath of solution-space libraries - which is where real gains productivity are achieved. Unfortunately many people don't understand great design (in any area) is generally about removing unnecessary complexity whenever they see it. As Einstein is often (mis-)quoted, "As simple as possible, but no simpler". Java is like C, it kicks ass because the language just enough to get the job done (and the real work is done by the libraries).

So then you're a believer that status quo is exactly the right balance between simplicity and power? What a remarkable coincidence

Re:They should not have used Java

[SplashMyBandit]

> Good, so we can agree then that, as far as getting old stuff working, WinForms is not any worse than Swing.

Assuming of course you never used pInvoke to get anything done (really, you didn't?) and I never used JNI (never needed to, unless I was integrating hardware).

> Also, I cannot claim any love (or hate) for SWT as a developer simply because I haven't written any app in it. All of my good attitude towards SWT and corresponding bad attitude towards Swing come from my experience as a user of apps written in either. That said, I'll take a Qt app over both any day.

Laughable. So you won't actually listen to a dev who's used these in anger and say they are a dog but instead only care the fact they that make you feel as if the were ctrl3d apps (and apps that don't look and feel exactly like this rate as worthless in your book). In the old days we used to call that arrogance. Here, be my guest, please take a mirror.

Look, I have nothing against lambdas or any other minor language feature. If I wanted to use them then Clojure would be a good option (out of many), and I may use them when they fall into Java. Turns out I find them mostly irrelevant for two reasons: I've been able to solve all my problems elegantly enough without them; and, they actually make negligible difference in projects when you're building for the million-lines and milions-of-users scale. The niceties of C# (for example, they it has definitely them) will *never* make up for the fact that you'll miss out on the business of Windows-focussed clients *and* the huge clients that have a mix of everything. For me it is about the business principles of making *money* and the continuity. Before you say it, there is no real productivity difference between Java and C# for a reasonable dev, so that really isn't reason to choose one over the other. One day .NET will be on the unloved heap along with OLE,COM,DCOM,COM+,ActiveX,MFC and all the other tech MS hyped at the time (and still allows to run, kinda). Just look at all the fools who backed Silverlight and how they are getting shafted now as Microsoft switches focus to a new technology (as they always do every few years). I bet you think it isn't going to happen to .NET, or that the cost of moving your (or worse, you clients) business over and re-writing everything is negligible (well, it's not, but then if you are only an employee you don't usually think about this stuff). I've been fortunately to have been around long enough in this game to see Microsoft do it again, and again, and again, and realise they *have* to do this to make money (at the expense of those that build systems based on their tech du-jour). I'm also smart enough to know that you can get away from the karmic cycle of continual tech re-birth and actually concentrate on solving hard problems instead of re-inventing the same sh!t in new tech. Sure I upskill continually (doesn't everyone?), doesn't mean I use the same fads to develop my critical stuff.

Oh, and if you are one of those people who only notice seismic shifts and never notice the creeping of tectonics (techtonics, lol), here's a question I have. Why would you hitch your horse solely to a stagnant company, with no alternative supplies of their tech, and one whose share price is only propped up by continual colossal share buybacks? Seems that the Wall Street dudes have a better grip on commercial potential than many, which is why although MS is very profitable (won't go away) it also not going anyway (if you have been paying very close attention you'll see it has started its decline). Microsoft is like IBM, huge, but no longer a thought leader. It seems a shame you can't even see that you are putting their business interests ahead of your own by limiting yourself to their tech - plus you miss out on all the big opportunities you are never even invited to bid for, since your choices have limited you so you can't do Windows *and* everything else. This is what differentiates tactical thinkers from strategic ones. It would be very interesting to have the same discussion ten years from now. Thanks for taking the time to post your reasoning and state your point.

Re:They should not have used Java

[shutdown -p now]

It would be very interesting to have the same discussion ten years from now. Thanks for taking the time to post your reasoning and state your point.

Indeed - let's agree to disagree and see how things work out a few years down the line. All it takes to resolve this dispute is a little patience.

By the way, my apologies for anything that can be perceived as a personal attack in my earlier comments in the thread. I do try to control myself and avoid such stuff when discussing things, but this being Slashdot and all, it still does slip through every now and then.

Oracle bought Sun for $7 billion

[airfoobar]

Looks like they want to make all their money back with this lawsuit. 'tis how they roll, I guess.

I see people saying Google should buy Oracle. Not gonna happen, of course. However, Google could buy the Java IP off them if they wanted. Surely, the valuation of Java's IP would be much smaller than the entire valuation of Sun (hw divisions and all), so it may actually come cheaper than fighting this out to the end.

On the other hand, that is probably what Oracle wants to happen. Tbh, I want to see Google fight this and rip Oracle apart in court, because Oracle are fucking asshats no matter how I look at this.

Re:Oracle bought Sun for $7 billion

[Anonymous Coward]

Looks like they want to make all their money back with this lawsuit. 'tis how they roll, I guess.

I see people saying Google should buy Oracle. Not gonna happen, of course. However, Google could buy the Java IP off them if they wanted. Surely, the valuation of Java's IP would be much smaller than the entire valuation of Sun (hw divisions and all), so it may actually come cheaper than fighting this out to the end.

Never worked much with Sun, I see.

If Java were worth $100 billion, Sun would have sold it for a dime.

On the other hand, that is probably what Oracle wants to happen. Tbh, I want to see Google fight this and rip Oracle apart in court, because Oracle are fucking asshats no matter how I look at this.

Yeah, but Oracle's the asshats that tend to win (and then rub it in). Java has a LOT of patents in it - look at how many hundreds of millions of dollars Microsoft sends each year to Sun-now-Oracle to license those patents for .Net.

Google played with fire by stepping outside the Java license to use those patents. IMO they were probably relying on the fact that Sun tended to be really, really nice about things like IP disputes.

Oracle's not.

IMO Oracle's going to win their billions of dollars from Google unless Google can somehow invalidate Oracle's patents.

Re:Oracle bought Sun for $7 billion

[Anonymous Coward]

People can say whatever they want, but the fact is that Google can't buy Oracle. Oracle is worth more than Google.

The C language

[Anonymous Coward]

Yeah seriously, I don't know why people even bother to learn languages that are 11 times as slow on fairly routine tasks. Yes object oriented stuff is nice, and functional languages do really interesting AI stuff, but you pretty much have to implement everything in C if you want to get it up to competition speed or if you want to make it highly scalable.

Ok I'm exaggerating a fair bit, but seriously C gets far too little love.

Re:The C language

[icebraining]

C gets far too little love

You're kidding, right? C is one of the most used languages, the second according to TIOBE (which is flawed, yes).

you pretty much have to implement everything in C (..) if you want to make it highly scalable.

That's so wrong I don't even know how to begin to tell you. Scalability has nothing to do with being fast per se, but with scaling linearly.

Re:The C language

[Just Some Guy]

Yeah seriously, I don't know why people even bother to learn languages that are 11 times as slow on fairly routine tasks.

Because I'm way more expensive than adding a second rackmount server. Because the few hundred thousand lines of Python in production in my company are either event-driven (waiting for web requests), IO bound (waiting on remote websites to answer), or scheduled (nightly batch processes). Because our "busy" server currently has a 15 minute load average of 0.13 and 99.2% free CPU. Tell me again why I'd want to write CPU-efficient code over programmer-efficient applications?

Re:The C language

[Lawrence_Bird]

Actually, Fortran gets far too little love. If anyone would bother to look at the more recent standards (say 95 and on) they would see that it does a heck of a lot (and very well too.) Unfortunately, people have been taught/told that it is either a) dead or b) evil.

Universal

[SnarfQuest]

Wasn't Java(tm) being forced upon programmers as a universal cross-platform programming language? Sun wanted Java everywhere. However, whenever I tried it as a cross platform solution, all I could get were errors of the type "Sorry, this program was compiled with java runtime of 1.4.5.4.3.4.1.3.2.33.4.5.p, and you are using 1.4.5.4.3.4.1.3.2.33.4.5.q" .

So, Google drank the tea, and now is being punished for it.

Re:Universal

[SplashMyBandit]

Back in the day I used to build my Java programs on Windows 95 and run them on 64-bit 4-CPU SGI IRIX at my university. These days I never had a problem with 'write once, run anywhere' when I compile on Mac and test in Linux, and then give my stuff to people using Windows XP 32, Vista, and Win7 32/64 and none of them have had a problem with my Java program (bundled as n exe with JSmooth). You must have been doing something wrong?

Re:Universal

[Anonymous Coward]

You must have been doing something wrong?

duh. he was using 1.4.5.4.3.4.1.3.2.33.4.5.p. That's like 3 days old man! That's like writing in QBasic!

Re:Universal

[umarekawaru]

I'm developing a multi-threaded music notation program in ancient Java 4.2 ( 1.4.2 ). I never had a problem running it unchanged on Windows XP, Mac OS 10.3.9 on PowerPC and Mac OS 10.4.11 on Intel. Playback graphics and sound are refreshed and can play 32nd notes on my ancient 867 Mhz Power Mac. Even my ancient version of the Java VM and language is amazing for development. I chose Java for the existing 3rd party libraries available where nothing else compares. I have one code base to maintain and I reach a vast user base including Linux users. I could not have done this alone had I used any other language or API.

Follow the money...

[Anonymous Coward]

Java made money for Sun (now Oracle) on mobile phones, it was licensed for use there, Google did something clever (or maybe not so much) and released a phone/framework that used a previously accepted implementation of Java (heavyweight versus the 'efficient' phone version) and skirted the licensing issue altogether.

Now Oracle feels that Google owes them licensing for 'phone' java.

So what is Android, a phone with java? or a tiny linux system with a phone card?

I believe this is one reason Apple wants to distance themselves from java, its going to become a headache.

Re:Follow the money...

[Anonymous Coward]

Well, my Galaxy Tab here ain't no phone...

The phones all have different processors

[Marrow]

Where is the glue language that makes a single distributable work on all platforms?
As far as Java goes, they should have come up with their own implementation/compiler whatever that does not draw from anything but the syntax of the language. The syntax is open right? And bytecode is not a new idea so they could have rolled their own clean-room bytecode without straining anything.

Re:The phones all have different processors

[Anonymous Coward]

They did roll their own clean-room bytecode. Dalvik does not use Java bytecode.

Beware "placed" articles

[davecb]

Groklaw also identified this as FUD, also known as "trying the case in the newspapers".

--dave

This is like watching Clemson play UGA

[EmagGeek]

I don't know who I want to lose more.

Business Decision

[blunte]

... spend 6 billion to make Larry Ellison's head bigger

or ... spend 10 million to remove Larry Ellison's head

From a pure business standpoint the decision is obvious.

Ah

[Greyfox]

That would explain why all my searches for Oracle on Google today keep redirecting me to goatse.cx.

Couple of errors there

[Estanislao Mart�nez]

First of all, cash isn't the only method for corporate acquisitions. The other one is that they buyer can trade their own equity for the purchase; "I'll give you n shares of my stock for each m shares of yours." This can be combined with cash, but cash is part of the buyer's market cap too, so to estimate one company's capacity to buy another, you look at the market cap, not at the cash reserves. GOOG are 155.99B, so they'd still have to give away more than half of their company to get half of Oracle.

Second, when a company is interested in buying a second one, they usually have to pay premium to convince the target's owners to sells. Buying Oracle would cost more than Oracle's market cap—or more precisely, a serious intent to buy Oracle would drive up Oracle's price.

Will They sue themself next?

[doctormetal]

The company that has done the most damage to Java is Oracle. Will they sue themself next? It seems that Oracle is trying hard to destroy all assets they acquired from the Sun.

Re:something other than Java

[TaoPhoenix]

What are the #3 4 5 languages? Aka count counting Java, not C++.

Let's say Google squashes this and everyone boycotts Oracle and Java. Anyone have thoughts on what the dark horse language is?

I keep hearing Python mentioned. Aside from the "whiter syntax" apparently it might let me write non-OO programs more like BASIC but more modern.

Oracle can go fuck itself.

[DoktorSeven]

Failed companies are pathetic trolls trying to steal money from companies who rightfully and successfully use what is available. Oracle, you're irrelevant now. Leave Java to the community, accept that OOo has been taken over, and let everything else you are hoarding go as well, and find some other thing to do.

When companies make such desperate acts to gain money and fame, they deserve neither.

Adobe?

[Anonymous Coward]

Next on Oracle fantasy island: Oracle owns ColdFusion and JRun because it's based on Java. Lawsuit pending.

oracle taking steps to destroy java

[Anonymous Coward]

Oracle's action of suing a company for it's use of java is a terrible action. If my business depended upon the use of Java I would take steps to move off java as quickly as possible. Oracle has showed a terrible disregard for the collective use of Technology.

sun was bought for 5.6 billion

[krups gusto]

how can the damages be more then the entire value of sun?

tax on bluster

[epine]

If there was even a 1% tax on proposed settlement size, the amount of bluster in these figures would collapse by an order of magnitude.

Imagine if it cost $63m to file a claim in the court system for $6.3b in damages.

We all know Ellison is going to ask for $2b/inch whenever he can get away with it, if there's no price for grasping.

Eating sushi, no ..

[ackthpt]

Know anything about Larry Ellison? He thinks he's a Samurai. You have underestimated his determination to bring Google to their knees.

Eating sushi does not make you a Samurai.

Building in Woodside, CA, a multi-million dollar estate, modeled on feudal Japanese architecture (i.e. no nails) and hand built by builders brought over from Japan to do the job, just might.

He's also been photographed wearing a kimono.

Probably has a suit of armor and blades in his office.

He was known at one time to be working out some form of management based upon Bushido - no word on any underlings committing seppuku.

Oracle the new Sco

[Anonymous Coward]

Sure looks it to me.

Re:something other than Java

[Kagetsuki]

I've coded in Python before, and it has quite a few issues. It's a decent language and it runs in a lot of places - but Ruby runs in all those same places and has a significantly larger array of really nice language features. Really popularity of a language isn't going to define how "good" it is, and more specifically how much better it will be for your application.

Re:something other than Java

[Anonymous Coward]

I've coded in Python before, and it has quite a few issues. It's a decent language and it runs in a lot of places - but Ruby runs in all those same places and has a significantly larger array of really nice language features.

That's the biggest problem with Ruby, in my eyes. To be blunt, it looks and feels like Python as reimplemented by Perl enthusiasts. You can typically make things "sort of work" quickly, but I've had a couple of projects where ditching Rails was deemed less of a bother than trying to solve the mystery of why our pages took ages to generate. After extensive profiling.

Well so-so

[LostMyBeaver]

Java isn't that bad of a language. But :
    - The libraries are not the best I've seen. I much prefer the layout of Qt or the .NET framework. Personal preference though.
    - Documentation after this much time should have been a great deal better. Java documentation is so-so and the library documentation is abominable.
    - The Java VM should have been completely revamped when it became obvious a hardware implementation was pretty close to impossible. It is relatively easy to accelerate Java by translating most instructions to another processor architecture, however, it's more of a compile step than an execution step. The Java VM really should be implementable in hardware.
    - The intermediate code format should store data as little-endian instead of big-endian. This is 2011. x86, x64, ARM dominate the processor world. Most other processor cores (such as TI DSP and MIPS) have the option of running in little endian mode. PowerPC and what's left of Sparc are the only big processors left which are big endian only. There are others like hitachi H8 and such, but I'm not even sure there are java virtual machines for them. Let's be realistic, file storage of large arrays of data should be little endian these days. And yet, Java still uses big endian. Time for that to change.

But I do agree with you... as a language, it's probably a better C++ than C++. I have personally been working on something that's half way in-between myself. Basically, C++ without code separation and without the need for forward declaration. It isn't a compiler per-say so much as a code separator. It converts #imports into #includes and converts classes into code + classes. Compile times have dropped considerably by forcing recompile of entire code bases at once.

Re:Well so-so

[DrXym]

Java libs do have cruft, as can be expected from their age - things like all the synchronized collections, StringBuffer, AWT etc. and from Java 7 the mess of date / time APIs has been superceded. I guess they would deprecate this stuff if they could but it's used too much to remove entirely. .NET and QT had the benefit of coming later and could learn from mistakes, but even .NET is gathering cruft, such as WinForms but at least it appears to be a more dynamic language.

There are piles of documentation & tutorials but to be honest the system APIs are the least of your troubles. Java becomes useful by virtue of the masses of 3rd party libs and learning those is 10x more effort, especially many of those like Spring, Hibernate, JUnit etc carry their own cruft.

The VM is okay and indeed there are hardware implementations, even some ARM processors have Jazelle which allows direct execution of bytecode. I think it would have been better register based but changing it is not a possibility. I'm not aware that Java is especially slower than .NET, they're about the same especially with modern VMs that would do JIT / Hotspot compilation anyway.

I think comparison between C++ the language and Java is only useful so far because C++ tries to be a superset of C whereas Java is a C-like ground up approach to a new language. I think if you wanted C++ like performance with a Java like syntax that something like Google Go would be worth checking out.

Google Helped Kill Sun

[Compaqt]

This. The whole Android kinda-sorta Java virtual machine kludge was just a hack to avoid paying Sun any money.

If they had come to a reasonable agreement with Sun paying them a couple hundred million for a catchall license, Sun could've still been in business. Of course, Sun had no business sense, either.

All valid points

[LostMyBeaver]

Yep... .NET and QT learned the most important lesson which was versioning. Nothing can't be fixed by deprecation of older technologies, even technologies like real CPUs get these benefits. It might have taken forever, but my latest Intel Core i7 doesn't even run 16-bit code unless emulated. Java was close to being the first of its kind... SmallTalk being a notable but much ignored predecessor (mostly because it was REALLY annoying to program with) and having set things in stone the way they did was a huge mistake in my opinion. I personally ported AWT to 5 different platforms back in the old days, even wrote a mostly Java implementation based on a frame buffer and an input event manager. I ported SWT to two platforms before Eclipse was even released within IBM for internal evaluation. And looking back on it, looking at AWT, SWT, Swing, I'm still amazed that Java is still used for making UI apps.

Let's forget the third party library system and focus on something more important. Naming.

Java, Mocha, Kaffe, Beans, Latte etc...

Don't get me wrong... I love coffee... drink over a liter of it by lunch time each day. But I have to Google to figure out what any of this crap is. If they'd have used three letter acronyms, I'd have had an easier time knowing what they were talking about. This utter lack of professionalism has always left me feeling like Java is the language of people who would rather be playing with tinker toys than writing code. What would have been wrong with something like "Object Modularity and Component System" as opposed to something stupid like beans.

Jazelle was always interesting because it took an acceleration approach as opposed to an implementation approach. In a way it said "Well, making a Java only processor is obviously a waste of time, but we can implement a means of running as many instructions as possible from directly within the CPU." I know there are hardware implementations of Java which accelerate probably as much as 90% of the system. And obviously, emulating 100% of the system would require an operating system to be implemented in byte code for it to be useful. It could be done I supposed, but it would be painful and ugly. As matter of fact, I remember some of the earlier attempts like JavaOS to do so. And I stand by the painful and ugly and add awkward to the mixture. PCI drivers were impressively ugly because of the weird method of memory mapping involved.

A more generic VM would have been much nicer I think. Start with a Motorola 68000 instruction set or even an ARM instruction set. Then compile the language natively to that core. Stack, register or Harvard disaster based, makes no difference. The point would be that the processor would be generic enough that the language and technology would be more important than the VM. These days when most CPU architectures include protections such as Intel's "Do not execute" flags and such, a more generic CPU with a specialized MMU for garbage collection would have been better. But, there's no point to it anymore.

The real value to Java is in the language itself. I don't know if I'd try to differentiate the relationship between Java to C or C++, but it's far enough from either and close enough to both that it would be hard to decide. It's a C/C++ like language as much as C# is. I personally program about 95% of my code in C++ since, with the exception of header/source mess which I no longer notice and by using Qt, I find it to be a great solution for everything from scripting to kernel development. Hell, it even has regular expressions :) I personally find that I can often write code in C# or Java that can perform equally well or better than in C or C++. Especially when handling large integers and large amounts of dynamic memory allocation.

VMs are really a thing of the past these days anyway. They're more closely related to being a second stage compiler. Of course, in a language like Java where some code really shouldn't be compiled unless that execution path is de

Java died when Oracle bought Sun

[wye43]

Who the hell will want to invest in a technology owned no longer by its creative creator who made it grow(Sun), but now by a patent troll(Oracle) who is only looking to sue the Java users?

RIP Java, my first true OOP love :(

Yes, I am a little confused about that.

[Marrow]

I thought dalvik converted java to something else, but its another tool that does that.

gmhowell, step inside

[Anonymous Coward]

http://slashdot.org/comments.pl?sid=2251620&cid=36512016