Oracle To Monetize Java VM

jtotheh writes "According to the Register, Oracle is going to make two tiers of Java Virtual Machine — a free one and a premium paid one. 'Adam Messinger, Oracle vice president of development, told QCon that Oracle plans to offer a "premium" edition of the JDK in addition to the open-source JDK. Both, it seems, will be based on a converged JRockit VM and the Hotspot JVM from Sun Microsystems. The converged JVM will be released under the OpenJDK project. ... Messinger didn't explain how the premium JVM would differ [from] the free version, but the premium edition will likely see performance tuning and tie-ins to Oracle's middleware.'"

Comments

mm

[chibiace]

the death of java?

Re:mm

[HelloKitty2]

No, this is the birth of new opportunities in the java landscape, this is a clear sign of Oracle's dedication to the java community. The high-end Mercedes offering will finally allow you to look down on those Fiat drivers and know that your money is well spent.

Re:mm

[Slackware95]

I agree with this statement, this is indeed the birth of new opportunity - A new technology to replace Java...

Re:mm

[levell]

Given that Google is on the sharp end of Oracle monetizing Java, anyone else think they might start to push Google Go really hard? It's immature at the moment but it looks really nice and I think as it matures it could really catch on.

Re:mm

[krischik]

Which part: Java, Java-VM, Java-SE, Java-EE or all of them?

Re:mm

[Anonymous Coward]

I'm personally hoping someone will come up with an open-source implementation of C# not based on the .NET libraries or the Mono toolkit, but a pure native-code compiler, with selectable manual or automatic memory management. I believe C# is 'better than the orignal' Java. It's only drawback is that it's tied to Micrsoft and Windows.

It's pretty obvious that Oracle is hell bent on either making bumper profits off Java or killing it. They won't have it any other way.

Re:mm

[HelloKitty2]

no classes, no game

Re:mm

[Lennie]

Only thing I know of is Vala.

Re:mm

[am 2k]

LLVM# comes to mind. Unfortunately, this is only a scientific project, not a community-based one.

Re:mm

[gru3hunt3r]

I agree - the fact Google uses/relies on Java is really an Achilles heel for them. I'm 100% certain that when Larry bought Sun he scratched a 10 year itch he's had for how "Sun ought to run their business" - by making two versions of java one "free" crippled, and another enterprise one. Get 'em hooked, then bend 'em over and make 'em take it up the ass and pay for the service - that is the Oracle business strategy isn't it? Ellison could care less if anybody new ever uses/deploys Java -- because the installed application base alone in fortune 500 companies (existing oracle customers) is easily enough to pay 10x for what Sun cost. It's a freaking mint, and I think he's doing the right thing (for his shareholders) - he's not interested in the long term business plan, only short term revenue.
Killing mysql (a competitor they were losing business to), killing open office was just icing on the cake, and monetizing Solaris were just a few of the ways he's planning to make money.

I imagine this conversation happening in the Oracle board room:
Ellison: "we gotta nip this free software thing in the bud boys, next thing you know our stupid customers will be expecting our stuff for free too." (look of disgust)
Henchmen #1: "yeah boss, but how we gonna pay for it? the shareholders will never buy it"
Ellison: "those morons at Sun have been doing it wrong for years boys, what have I always told you"
Henchmen #1: "the customer will always pay more?"
Henchmen #2: "who cares if it's crap, ship it anyway?"
Henchmen #1: "who cares if my jet wakes people up? i'm rich?"
Henchmen #2: "nothing is sweeter than making the customer pay up the ass for crap?"
Ellison: "no, well - yes, I've said all those things, but I'm talking about how I'd run Sun, how I'd make everybody pay for Java, nobody should expect to use it for free"
Henchmen #2: "oh yeah boss, that was a good one"
Ellison: "look at this boys, it's like it's a god damn christmas - we stop mysql for a few years while the community 'forks' or whatever, you realize how much revenue that is going to protect for us?"
Henchmen #2: "oh yeah boss, that's alotta money"
Ellison: "then we kill open office, teach anybody who bought it a lesson, nothing is free - you want to use it - you should pay for it"
Henchmen #2: "yeah boss, keep going"
Ellison: "you realize how many of our customers depend on Solaris - they can't replace it for at least a few years, in the meantime we can tear them a new asshole and let the money flow out"
Henchmen #2: "that makes sense"
Ellison: "and then there's Java, wow.. what a stupid bunch of dumbfucks Sun was, I'll replace their free love society with Larrys pleasure palace where you have to pay me for some action"
Henchmen #2: "you mean metaphorically right boss?"
Ellison: "hard to say, all i know is that in the next few years boys, we're definitely going to be busy screwing all Suns customers up the ass, and charging them for the pleasure of it"
Henchmen #1: "so you mean basically we're going to do business as usual here at Oracle Co.?"
Ellison: "exactly"

The key word in business is "momentum" - the Sun acquisition took momentum from so many projects, and anybody that was using those projects (for commercial purposes) now is in the unenviable position that they need to either starting pay Oracle, or try and find a viable competitor (at least 5 years). In the short term everybody will pay, do you realize how many billions of dollars we're talking about - in 5 years they'll wash rinse repeat. This is the cycle we should expect to see in the future - I think it will be very good for Oracle (bad for the community, but nobody really gives a damn what those free-loving hippies think anyway)

Remember Fortune 500 CIO's can't risk their enterprise to free "crippled" versions of software, they can't use unproven forks, if something goes wrong - it's their ass (and bye bye stock options), they'll choose free only when they absolutely have to. Nobody cares how much money "they save", it's a corporation, it's not about saving, it's about CYA.

Oh my god I hope the folks at Oracle never get ahold of ASF.

I have to admit - the folks at Oracle are brilliant (from a shareholder perspective) because they get how big businesses work.

Re:mm

[kiddygrinder]

you should get that written on your pants, you know, for the irony

Re:mm

[aliquis]

I agree with this statement, this is indeed the birth of new opportunity - A new technology to replace Java...

You know a product is dead when even Microsoft offering is better.

Re:mm

[Tanktalus]

Well, I think Java-the-language is already out there. I can't imagine a way for Oracle to get those worms back in the can. So we must be talking about the VM, and maybe their extensions (SE/EE).

My prediction? The VM continues to be free. EE becomes premium. SE, not sure about. And there may be a pay-for-fixes model for the VM that is, by definition, not free. If you need something fixed NOW, you pay. If you need more performance from the VM, you pay. Those fixes may or may not make it into the generally-available VM.

The community solution? IMO, it is to finish Parrot, get a "Java-the-language"-to-Parrot compiler built (perhaps by starting with a Java-the-VM-to-Parrot bytecode converter), and then you can pretty much discard Java-the-VM. Parrot already supports a bunch of different languages (at different levels of completeness), so it seems to me to be a natural fit.

After that project gets going, approach gcc to have gj able to target Parrot instead of Java bytecode, like all their other cross-compiling solutions. It makes Oracle irrelevant.

Maybe, though, we'll have to wait and see what Oracle's real plans are before we, as a community, start down this road. It's an expensive road (in manpower more than $$), and if everyone continues to be happy with Oracle's free offerings, there won't be much impetus to go around Oracle. That said, I can imagine some people still worried enough about Oracle's next move to start work on this, or a similar, solution.

Re:mm

[dynamo]

So true. mod parent up.

Re:mm

[c0d3g33k]

You already have that in Mono. Mono is fully open-source/free/libre, there is no obligation to use the .NET libraries - you can ignore them entirely, and Mono can do full AOT (ahead of time) compilation to native code already. I'm not sure what else you want exactly that doesn't already exist.

Re:mm

[c0d3g33k]

I'd like to hear about some serious deployments of Parrot in production before I'd even consider it a candidate for a viable alternative to the JVM. Maybe it's further along than I realize, but since you state that it still needs to be 'finished', it's not a viable contender in the context of this discussion.

Re:mm

[makomk]

This seems like a good incentive to stick with open source projects that aren't tied to a single corporate owner, like PostgresQL. Companies still can and do offer commercial support for Postgres - in fact, some of them have the advantage of employing a number of the key developers - but there's no single company that can be targeted for a buy-up like this. (While Oracle could tray and buy all of the companies offering support for a particular option, one would hope that this would attract the attention of monopolies regulators.)

Re:mm

[mario_grgic]

Both funny and insightful. So true.

Re:mm

[mario_grgic]

LLVM is not just a scientific project. Apple is seriously pushing for LLVM and they intend to use it as default compiler in OS X.

Re:mm

[mario_grgic]

I thought Parrot was geared more towards optimizing performance of dynamic languages, and not so much statically typed languages like Java?

Re:mm

[am 2k]

I was talking about LLVM#, the C# compiler for LLVM, not LLVM itself.

Re:mm

[jon3k]

Wow, someone mod this guy up. Excuse me while I take my first serious look into Mono.

(not sure if that sounded sarcastic - but I'm being serious I had no idea)

Re:mm

[ewanm89]

class Main {
public static void main(String[] args) {
System.out.println("Hello World!");
}
}

I count a 5 line hello world that he is suggesting replacing. Striping out extrenuous whitespace characters gives 65 characters for google go and 85 for the equivilent hello world in java. Now I'm not suggested google go is better or worse than java, in fact this is the fist I've looked at this particular language. Just that first argument is useless in context.

Re:mm

[TheLink]

Parrot? Is that in the resting on pining for the fjords stage?

At the rate things are going, Javascript has an even better chance become the next thing than Parrot. Seems like everyone and their dog is obsessed with making it faster.

Javascript is not as fast as Java but...

http://shootout.alioth.debian.org/u32/benchmark.php?test=all&lang=v8&lang2=java

Seems it's actually not that slow unless you're trying to do stuff like calculate digits of pi.

In contrast stuff like Perl, Python and Ruby don't appear to be getting faster as fast. Javascript is already faster than all these 3. Go figure what the "community" has more interest in.

Don't get me wrong, I'd be happy if Parrot, Perl, Python and Ruby were all much faster (and complete in Parrot's case), but I'm not going to hold my breath.

Re:mm

[Nelson]

It's a certainty. Thing is, google-go is a C competitor, not a C#, Java, etc.. competitor. Not any time soon at least and it's also very much a work in progress.

Sun beat MS in a very similar law suit, I don't see any reasons why Oracle won't win against Google either, that leave google 2 options: 1) buy the rights from Oracle for a huge chunk of cash or 2) create their own tool set and language. They've already built the tool set.

Personally, I see it as a marketing and product positioning problem. There is clearly a market for a C competitor, people have been working at them for years but no real players have backed any until go. Google pretty clearly needs a higher level application tool too. They just need to market both without eating each other or undermining each other, that's the problem.

Re:mm

[LukeWebber]

Bingo. If Go becomes the dominant language, I'll be all the happier that I've jumped ship to C#/.NET. Ugliest damned labguage I've seen in a long time.

Re:mm

[owlstead]

I've tried it and got into it. I gave up once I found out that they somehow think that exception handling shouldn't be a first citizen. I mean, building a *monolithic* API loosely based on Java without exception handling? They are as mad as the name implies, it's about the only non-googleable name in the language multi-verse.

Re:mm

[TheNarrator]

What about the LLVM Java JIT? It apparently runs Eclipse and Tomcat.

Re:mm

[shutdown -p now]

I'm personally hoping someone will come up with an open-source implementation of C# not based on the .NET libraries or the Mono toolkit, but a pure native-code compiler, with selectable manual or automatic memory management.

Your requirements are contradictory.

With respect to "selectable manual or automatic memory management" - that would require an extension of the language, since, as defined by the spec, it assumes automatic memory management. All of the existing code, including the standard libraries, assumes so as well, so you cannot reuse it. Heck, even the standard library APIs kinda assume it, since they often do not specify whether transfer of ownership occurs or not where it would be important with automatic memory management.

Pure native-code compiler is doable while remaining strictly to the letter of C# spec, but as soon as you wade into the neighboring CLR realm, you'll need either a bytecode interpreter or a JIT-compiler - due to existence of things such as System.Reflection.Emit which enable on-the-fly code generation - and which are used in existing code.

Anyway, if you want a new language which takes quite a lot of good ideas from C# (including a fair bit of syntax), but which is native-compiled, produces libraries which are directly consumable from C, and for which Unix is a native environment - look at Vala.

Re:mm

[shutdown -p now]

My prediction? The VM continues to be free. EE becomes premium.

EE (as well as the rest of Java) is just a set of spec. There's nothing preventing other vendors from offering free implementations of it. In fact, that's precisely how it worked for most of Java's history - you could buy an all-in-one stack from one of the commercial vendors, or assemble it from FOSS offerings. It is only recently that Sun itself started to offer a complete free implementation.

Re:mm

[shutdown -p now]

I agree - the fact Google uses/relies on Java is really an Achilles heel for them.

Why? OpenJDK is still FOSS. If Oracle starts to push their own agenda, Google (and likely other interested parties - IBM?) would just fork it. The only catch is that they wouldn't be able to call it Java due to trademarks, but who cares, especially for internal use?

Re:mm

[perlchild]

Presumably all of them, starting with Java-EE of course, but and this is the tragedy...

Bye bye JavaME

Re:mm

[oldhack]

You don't know Java.

EE is bunch of specs, not a part of language to be implemented in the VM.

Perl people (Parrot) would be the last ones, and by that I mean HELL NO, to produce anything as useful and comprehensive as the worst of the Java universe, including the god-awful JEB idiocy. At least the JEB idiots got it out the door.

I hope Gosling would start up a project to continue Java lifeline. If there is a single software project that can measure up to Linux, it would be Java.

Re:mm

[chromatic]

It's possible to do both, especially when you realize that optimizing a dynamic language well means understanding the 90% of the program you can optimize statically and doing so.

Re:mm

[chromatic]

At least the JEB idiots got it out the door.

Permit me to introduce you to the CPAN.

Re:mm

[oldhack]

You can introduce frying pan for all I care. Where the hell is release quality Parrot?

Re:mm

[Qubit]

If Go becomes the dominant language, I'll be all the happier that I've jumped ship to C#/.NET. Ugliest damned labguage I've seen in a long time.

If this is what programming in C#/.NET does to a man, then count me out.

Re:mm

[CarbonShell]

IIRC the 'free' is only by virtue of M$ and lives at the whim of M$.
The last company I'd trust to keep it's word.

Re:mm

[CarbonShell]

As long as it can offer the features Java has and is fully free, great!

I have done some work with Vala and while it is nice, does not go far enough.
But who knows, with some more focus the libs might be extended and we could see some improvements.
Something like a classpath would be worth gold.

Re:mm

[daem0n1x]

Messinger didn't explain how the premium JVM would differ [from] the free version, but the premium edition will likely see performance tuning and tie-ins to Oracle's middleware.'"

Just what I need, tie-ins. I'm even willing to pay for them.

Re:mm

[gru3hunt3r]

PostgresSQL has a fraction of the momentum that Mysql has, or at least had.

By Oracle literally buying and slaughtering their leading open source competitor they effectively showed that Goliath, equipped with an expensive helmet would have absolutely routed David.

For big companies (the type who pay for support to fund those projects) they won't put their eggs into a basket if they think they are likely to get crushed. Oracle is better off waiting till a project gets some momentum then killing it. Each battle will be different, but because of Oracle's war-chest of money they will ultimately win.

Get used to it.

Re:mm

[mcvos]

Yeeech tab-indent or get back to the 20th century.

You're kidding, right? How many modern languages actually do that? Most people are not convinced that blocks-by-indentation is an improvement.

Re:mm

[JAlexoi]

If they try to aggressively monetise Java, then they will be picking their fight with IBM. And Oracle doesn't want to do that under any circumstances.

Re:mm

[chromatic]

Parrot's 22nd release since 1.0 comes out next week.

Good.

[Anonymous Coward]

Good, now maybe someone will come up with a better language to use.

not trolling or anything, the only reason java is tolerable at all these days is because finally, most devices, including mobile devices, have become powerful enough to withstand the sheer bloat. 10 years ago, running anything java was painful.

now modern mobile devices are more powerful than those systems from 10 years ago, I think that's the only reason java is usable.

wanna bet the premium version will promote cross-platform compatibility while the free version is restricted to whatever platform oracle chooses? (ie, more than likely windows)

Re:Good.

[devbox]

There is already C#. Compared to Java, it has been the better language all time.

Re:Good.

[Bert64]

And now that hardware has finally caught up with the bloat of java, people are moving to things like ruby which are even slower...

Re:Good.

[EsbenMoseHansen]

Hopefully, noone is using ruby because of the speed ;) If you want speed, there is C++ and friends.

Re:Good.

[kiddygrinder]

you're an idiot if you use ruby for speed, you're just as dumb if you think ruby is bad because it's slow

Re:Good.

[SanityInAnarchy]

Well, there's also the fact that Java performance has caught up to the point where it's not quite twice as slow as C++, and faster in pathological cases. And there's the fact that Java, the language, isn't what's in question here -- it's the JVM, which means JRuby, Scala, Clojure, etc.

So it's not just 10 years of Moore's Law, it's 10 years of optimization. Java was pathetically slow. Now it's the fastest thing in its class.

So what would you replace it with? C# isn't necessarily faster, and it's got that wonderful Microsoft lock-in. Anything else is either going to be much slower (perl, python, Ruby) or much more dangerous (C, C++). It's possible there are some obscure Lisps that come close, but we're at the point where all the major optimization research goes into either C/C++ or Java, to the point where CPUs are designed to run C fast, not the other way around.

Oh, and I expect it'll be the other way around -- the free version will be cross-platform, while the premium version will be "vertically integrated" -- might be one Windows implementation, but very likely one more, maybe on Solaris, maybe on "Oracle Unbreakable Linux", but somewhere they can "integrate" it more thoroughly into the system. At least, that'd be the typical move for them, and the smart move -- just another way for them to grab the enterprise by the balls and squeeze, while ignoring everyone else, especially their smaller customers.

Re:Good.

[otis wildflower]

I'm actually looking forward to Perl6, but it seems to me the decision to leave implementation up to "the cloud" rather than have a complete reference implementation is delaying it and may destroy it. I don't want to end up with 5 different implementations that each support "a bit" of the spec, and that looks like where Perl6 is right now, unfortunately.

If there is implementation forking, it should not be on different areas of the base language, it should be in different optimization choices. Perhaps a 'slimmed down' optimization would be OK IFF there's a fully-implemented version available for the general case.

Re:Good.

[RightwingNutjob]

Anything else is either going to be much slower (perl, python, Ruby) or much more dangerous (C, C++).

You know, I never quite figured out what this 'dangerous' thing meant when talking about C or C++. Yes, there's nothing stopping you from going *(int*)NULL = 1337;, but programmers are supposed to be intelligent detail-oriented people who after a few years of training can instinctively and deliberately avoid bugs both the subtle and the egregious without wasting time that is supposedly saved by higher-level languages like Java that do everything for you including tying your shoes for you in the morning.

Re:Good.

[ADRA]

"intelligent detail-oriented people who after a few years of training"
I guess that eliminates most of the programmers out there... so sad.

Yes, because every programmer on earth is just as you describe. They are perfect beings that NEED to use pointers in order to get their dull little boiler plate web application written. Oh wait, you mean pointers don't serve a specifically beneficial role in programs based purely on 99% of business domain solutions? That's insane! They're so elegant, they have to be woven into every programmer's brain. I mean how else can you write the most tight elegant sexy for loop if you're not using array pointer iteration.

Yes, its all hyperbole and so is your statement. C/C++ mandates a large set of things that MUST be done to get a job done. Some jobs are much better served if you use pointers, and macros and all the archaic and low level sugar that these languages support. Most programmer these days don't need such a level of control over the innards of the program in order to make an ideal tool, and in fact, moving many existing applications into C/C++ may incrementally improve performance while causing all sorts of problems in accessibility, bugs, and portability.

Another thing that I think would suffer is the rich and wonderful ecosystem of Java libraries that have grown over the years. I'd hate to have to glue libraries together in the classic C standard ways. In fact, I would find it as a big step backward. Maybe the .NET folks have come up with something more compelling than classical approaches, but who knows.

Re:Good.

[SanityInAnarchy]

Here's the problem: Intelligent, detail-oriented people make mistakes.

after a few years of training can instinctively and deliberately avoid bugs both the subtle and the egregious...

If you can find a single programmer who actually does that, post their "flawless" code along with a sufficiently-motivating bounty and see how long it lasts. To be fair, you should have some way to verify the amount of time it took to produce this code.

In the real world...

Wait, back up.

higher-level languages like Java that do everything for you including tying your shoes for you...

Wow. You're actually claiming Java ties your shoes for you? Java, the only language I know of where == can't be counted on for equality (because Operator Overloading is Bad, mmkay?), where primitives are special cases, where null is a special case, where threading is still handled via the same primitives as in languages like C or C++... That language?

Seriously?

I'm just going to pretend you didn't say that, so I can pretend you had an intelligent point worth my time to respond.

So, in the real world, where Java is only moderately higher-level than C/C++, programmers can and do make mistakes occasionally. Even the best occasionally make stupid mistakes. Just off the top of my head:

struct foo * createFoo() {
  struct foo val;
  val.a = 1;
  val.b = 2;
  return &val;
}

Yeah, I know it's stupid, first-year mistakes. But think about the concepts you had to summon up to tell me why that's wrong. And of course, if I do it this way:

struct foo * createFoo() {
  struct foo *val = malloc(sizeof(struct foo));
  val->a = 1;
  val->b = 2;
  return val;
}

Now the caller is responsible for cleaning up after me. Sure, I can create a function that helps, if foo had a bunch of additional stuff that needs to be individually free'd or otherwise released, but they still need to call that. If they don't, I leak memory.

We can do that, or we can do C++, where we get gems like this:

class B: public A ...
A a;
B b;
a = b; // whoops, the contents of b just got sliced!

Or if I decide to use the heap...

A *a = new A();
a = new B(); // whoops, I just leaked the old value of a!

Even if you do manage to do it perfectly every time, you're still spending far more time -- even the sheer amount of typing saved not having to deal with this bullshit is significant. And you won't do it perfectly every time.

And when you screw it up in C/C++, you leak memory, segfault, corrupt yourself, or introduce a security vulnerability. When you screw it up in Java, so long as you're not using threads, about the worst you do is a null pointer exception.

Yes, you can leak memory, corrupt yourself, or be insecure in Java, but an entire class of bugs are now not possible. It is no longer possible to segfault, and it is no longer possible to introduce security vulnerabilities or leak memory through your use of pointers or references.

And you know what? For 99% of what I do with a computer, I'll gladly take a 50% performance hit for fewer bugs. In fact, I'll gladly take a 95% performance hit (and I routinely do) for even fewer bugs and (much) faster development. It's just more important for it to be reliable and maintainable than it is to satisfy someone's ego about how smart they are that they can use pointers.

Re:Good.

[cowdung]

According to a recent talk I saw in www.clei2010.org (a peer reviewed conference in Latin America) the Java/C gap is now down to an average of 20%.

Re:Good.

[yuhong]

not trolling or anything, the only reason java is tolerable at all these days is because finally, most devices, including mobile devices, have become powerful enough to withstand the sheer bloat. 10 years ago, running anything java was painful.

Yep, that is exactly why Sun created Java ME. And they were going to monetize it, hence the Apache Harmony test kit field of use disputes. In the end, it was beaten by Android using a I think language-compatible but not bytecode-compatible clone of Java called Davlik, which killed it, hence why Oracle is suing Google.

Re:Good.

[vux984]

So what would you replace it with? C# isn't necessarily faster, and it's got that wonderful Microsoft lock-in.

At this point C# and the CLR has quite a bit less Microsoft lock in than Java and its Oracle lock in.

Anything else is just FUD. The licenses around C# and mono are really pretty good. Mono is GPL2/LGPL2/MIT and gives the C# compiler and CLR.

C# and CLI are ISO and ECMA standards with RAND licensing. That's about as good as it gets really.

I concede there are absolutely some patent risks with it but I challenge you to name something without patent risks.

You could clean room your own language runtime devise your own language... and still be sued into oblivion for patent infringment by Microsoft, Oracle, and probably IBM too... hell... you don't even need to have actually infringed any patents... SCO is the textbook example.

Re:Good.

[vux984]

You know, I never quite figured out what this 'dangerous' thing meant when talking about C or C++. Yes, there's nothing stopping you from going *(int*)NULL = 1337;, but programmers are supposed to be intelligent detail-oriented people who after a few years of training can instinctively and deliberately avoid bugs both the subtle and the egregious without wasting time that is supposedly saved by higher-level languages like Java that do everything for you including tying your shoes for you in the morning.

One can write in assembler too. We use C/C++ because we can focus on higher level problems without having to focus on all the minutia of assembler. We use C#/Java because we can focus on higher level problems without having to focus on all the minutia still exposed by C/C++.

Being intelligent and detail oriented still involves spending considerable time taking care of those details. Using a language without those aspects means you can spend your time being more productive.

Re:Good.

[SanityInAnarchy]

At this point C# and the CLR has quite a bit less Microsoft lock in than Java and its Oracle lock in.

How so?

The licenses around C# and mono are really pretty good.

Better than the licenses around OpenJDK?

Mono is GPL2/LGPL2/MIT and gives the C# compiler and CLR.

It's also incomplete compared to mainstream .NET. OpenJDK has the full 15 years of research and optimizations piled into the JVM, and so far, we get pretty much the same JVM they do.

While the CLR may be theoretically better at handling other languages, the JVM has the benefit of multiple production-ready languages targeting it -- Java, Scala, Clojure, and JRuby come to mind, and I don't know of four production-ready CLR languages.

C# and CLI are ISO and ECMA standards with RAND licensing. That's about as good as it gets really.

It's close, but note that "RAND licensing" says nothing about licensing cost, particularly licensing cost with respect to patents. So that puts .NET in exactly the same place Java is -- either Microsoft or Oracle could suddenly decide to start charging for the patents required for either platform.

I challenge you to name something without patent risks.

It helps when a platform has been around for 15 years without its creators ever trying to patent anything about it -- though that doesn't apply to Java.

In more practical terms, the preferred platform for .NET development is Visual Studio. The preferred platform for Java development is Eclipse.

So please, exactly where can Oracle lock Java in that Microsoft cannot lock .NET in? The only thing .NET has in its favor is Mono, which isn't quite equivalent -- and Java has its own share of not-quite-equivalent things, too, like Kaffe and gcj.

And all this is academic unless there's a reason to consider .NET in the first place.

Re:Good.

[master_p]

So what would you replace it with?

Ada? it's as efficient as c++ and as safe as an imperative language can be.

Re:Good.

[shutdown -p now]

It's too safe for "this has to be done yesterday" approach to software project management that is so prevalent today (in Java shops as well).

Re:Good.

[chromatic]

... it seems to me the decision to leave implementation up to "the cloud" rather than have a complete reference implementation is delaying it and may destroy it.

On the contrary, that's the best way to create a usable specification.

Re:Good.

[Tim C]

but programmers are supposed to be intelligent detail-oriented people who after a few years of training can instinctively and deliberately avoid bugs both the subtle and the egregious

I like to think I'm reasonably intelligent, I've been told I'm detail-oriented, I have a few years training and 11 years commercial experience building on a few years worth of hobbyist experience in a wide variety of languages, though most of those 11 years have focussed on Java and C#, with a couple of years of C/C++ thrown in at the start.

Somehow, I still fail to manage to write 100% bug-free code, as does every other programmer I've ever worked with.

There is no silver bullet, and that includes just being that damn ninja.

Re:Good.

[Peter Amstutz]

1998 called. It wants its flamewar back.

One of the primary reasons manual memory management sucks is even if your code is beautiful and perfect and doesn't let a single byte slip by, you might be required to interface with a library written by a neanderthal that codes by beating his club on the keyboard and creates all sorts of incidental memory errors that don't affect the immediate functioning of the library, but have side effects for your program as a whole (memory leaks especially). Memory safe, garbage collected languages dramatically reduce these sorts of problems.

Re:Good.

[SanityInAnarchy]

And yet, it manages to make Java syntax look elegant.

with Ada.Text_IO;
 
procedure Hello is
begin
  Ada.Text_IO.Put_Line("Hello, world!");
end Hello;

end Hello? You've gotta be kidding me!

Re:Good.

[SanityInAnarchy]

Outputting C is generally a bad idea, even if the rest of the language is a good idea. There's a reason that while technically every valid C program should be a valid C++ program, we still have separate C and C++ compilers, and C++ compilers don't output C.

But as another poster clarified, the performance gap is down to something like 20% between Java and C, and Java will give me things I don't think your choices can, like actual runtime reflection and optimizations. Also, a quick glance at Eiffel shows it has some sort of garbage collection, so is it really as fast as C?

At this point, unless it looks like Oracle is actually going to kill Java, I'll take a high-level JVM language over a high-level compiled-to-C language any day. JRuby is actually getting pretty insanely fast.

Re:Good.

[mcvos]

Not to mention that C++ isn't much of a replacement for Ruby. I believe Scala is the accepted alternative for people who need something faster than Ruby.

Suicide? The end of java.

[WolphFang]

Suicide? Sounds they are working on ending Java in a hurry. :(

Re:Suicide? The end of java.

[Anonymous Coward]

Pretty much, yes. At least, the end of themselves as an open-source company. It remains to be seen if their special sauce will be worth buying. It probably will be, for people who are already trapped into Oracle/Sun (just as it is with Intel's compilers, that just happen to smoke everyone else's on Intel hardware) but for everyone else the world will just go on as it has.

Re:Suicide? The end of java.

[MrCoke]

Why is this any different as Qt (also has 2 tiers)?

Re:Suicide? The end of java.

[WolphFang]

User base. Psychology. That which is open is now closed. And by the way 'f*' off to the community in general. I am sure more than one group of developers is seriously looking at core language jump.

Re:Suicide? The end of java.

[whiteboy86]

... people buying.. who are already trapped into Oracle/Sun...

And this is children why Java, Obj-C, C# and other 'corporate' languages exist in the first place.

Re:Suicide? The end of java.

[VendettaMF]

>> I am sure more than one group of developers is
>> seriously looking at core language jump.

I know we are.
Oracle is quite simply not a rational enough company to entrust with our ongoing development.

Decisions, decisions.C# on MONO or C++ as itself. Pros and Cons for each are obvious, but the actual final scores will be hard to determine efficiently.

Individually I'm leaning to C++ and arguing the point that sooner or later MS will do to C# & associated platforms what Oracle is doing to Java now, and what lawyers may say about MONO now will count for nothing in reality when MS decides it's time to drop the hammer.

Re:Suicide? The end of java.

[modmans2ndcoming]

name a single time MS has ever done that with one of their languages.

C# has been around for almost 11 years now. Seems the language is well out of its infancy. You are just spreading hate on c# because it is an MS creation.

Re:Suicide? The end of java.

[modmans2ndcoming]

BTW... mono allows you to use monotouch so you have access to the iOS API in C#. it will only be a matter of time before they expand that to include Android. One platform to target 3 Smartphone OSs? Sound like a good deal.

Re:Suicide? The end of java.

[bjourne]

Money is very important to a company. Remember, the only reason why Oracle controls Java instead of Sun is because the latter totally failed to make any money on their language, despite it dominating the computer world.

Re:Suicide? The end of java.

[gbjbaanb]

Considering that C# is Java with a fair few mods. I'm ambivalent whether that makes it substantially better than Java or just another Java, but the point is that once Oracle starts succeeding here, Microsoft and the new board might decide to take note.

Whilst Microsoft has never traditionally done this kind of thing, they used to be run by engineers (sort of, you get where I'm coming from) where they wanted to try and achieve technical excellence, they are now run by businessmen who are only interested in the share price (and their bonuses). Note that Microsoft heavily licences java to use in C# (Gosling said they pay Sun, now Oracle 'tasty' fees very year).

The only thing that might stop them from tier-ing .net is that they use it to sell other Microsoft software - you don't write C# code in notepad for example (try it! its hurts) or run it on anything other than Windows. Whilst you can do this today doesn't mean they won't crack down on it in the future - and you can so easily imagine them doing so.

At the moment, only the truly open languages are safe to invest your time in. All the company-owned ones are poised to be monetised at the drop of a hat.

Re:Suicide? The end of java.

[obarthelemy]

I think people are confusing language and platform. Say you decide to go to C++. Now you've got to choose a compiler, libraries, toolsets, editors, CVS... that's another bunch of people to entrust with your hard work

Re:Suicide? The end of java.

[VendettaMF]

I've been using C# for windows and iPhone development for 2 years now.

But the long and the short of it is that it is a non-free non-open collection of code, libraries and other binaries and there is a definite risk attached to using it in ways that do not directly benefit Microsoft.

Re:Suicide? The end of java.

[VendettaMF]

Thank you. That is the point I wanted to convey, and exactly my stance and viewpoint on the matter.

And phrased and explained a lot more eloquently than I managed.

Re:Suicide? The end of java.

[turkeyfish]

There is an obvious trade off. One could not expect ubiquity if your "product" is not either free or very low cost. Yet that implies less revenue.

Oracle's strategy could work very well, where Sun's has failed, by providing two parallel tracks. One a premium track and two an open track. The premium track would essentially put paying customers perhaps a year or two ahead of the community open track users, who might see these features in the free "community" track a few years later, after something better is in place in the premium track. From a business perspective users would be happy to pay if it gives them a business advantage. From a community perspective users would be happy as they could rely on essentially free code that they could incorporate and tweak to their own ends, acknowledging that free doesn't always mean access to the most "valuable" code in the market place of ideas and corporate balance sheets.

For Oracle to succeed it need only better refine Sun's model. They really don't have to kill Java as a free language, for on the language side they are in a position to differentiate between a free and a premium track on the JVM. The difficulty will be that they will need to bear more of the burden for developing the JVM on the universe of platforms they wish to see a "write once run anywhere" Java t actually work on.

As a former Sun investor, who lost BIG TIME for Sun's failure to grasp a viable strategy of how to monetize Java (not that I was expecting to make a killing in the first place), I wish Oracle well in the goal of making Java succeed. The Java concept is still a worthy goal, but it will only work if the "write once, run anywhere" philosophy doesn't disappear at its core. If that happens, Java will become highly fragmented and its fate and influence in the market (open source + private source) going forward will slowly diminish. If they handle it badly, Java's lock on corporate computing might be pretty short lived and perhaps many of Oracle's products that are now intertwined with it as well.

As always, customers will buy into a software-platform, if they believe it is in their interest to do so. This will vary as not all customer interests are aligned. Oracle has the option of choosing just how much of a differential it will make between "premium" and "community". They may make more money over the short term if they set the difference to great but loose in the long run, if they drive the community elsewhere. It will be interesting to see what happens or just where within the many elements of the Java universe, the differences are placed.

Its not as if others haven't made mistakes with Java. Sun erred in failing to establish a clear two tiered strategy and then late in the game demanding too much of some of its bigger customers who became too dependent on Java, such as Google, who went off to develop their own "JVM" like construct (whether its different enough from the original Java is for the courts to decide). In the long run, if Oracle can keep the difference small, it will capture a larger share of the entire market. If not, expect numerous forks to appear in the future as they are beginning to do so now as Oracle tries to figure out if it is extending an open hand to the Java community or just a fist.

Re:Suicide? The end of java.

[Anonymous Coward]

... people buying.. who are already trapped into Oracle/Sun...

And this is children why Java, Obj-C, C# and other 'corporate' languages exist in the first place.

One of these things is not like the other.
One of these things just doesn't belong.
Can you tell me which thing is not like the other
before I finish this song?

Re:Suicide? The end of java.

[KingMotley]

Microsoft doesn't/didn't pay Sun for licensing java for c#, they have had some licensing issues with J#, but the whole point of C# was to get away from Suns licenses.

You may not write (or be able to write) C# in notepad, but many can and do quite well. The same can be said about any language in which there already exists a good integrated development environment.

I agree, there is sometimes merit in using open languages. I would suggest starting with ISO/IEC 23270:2006. Oh wait, that's C#.

Re:Suicide? The end of java.

[RocketRabbit]

When has Oracle ever been an "open source company?"

They will use free geek help when it contributes to their bottom line. They are not interested in making a shiny happy world for poor children in Burma. Do you think Oracle contributes to Linux because they want to feel good about themselves?

They do it because Linux is widely known, and Linux admins are a commodity. There's a lesson here somewhere, for those that pay attention.

Re:Suicide? The end of java.

[tibit]

The difference is that with Qt you at least used to get your money's worth -- there was way to get bugs mostly fixed, features got implemented, you could get real help. You pretty much knew what you were paying for. And they have this clause that any code developed under GPL cannot be transferred over to commercial license; some argue that it's unenforceable but that's what you have.

Good luck getting any of that from a behemoth the size of Oracle. Paying for Java is merely a way of corporate CYA, it's useless otherwise.

Re:Suicide? The end of java.

[tibit]

IMHO Java needed to have been spun off long time ago into a small company whose core focus would be just Java. Pretty much duplicating Trolltech's focus, dedication and agility. Not only would it be self-funding, I'd be quite happy to pay for Java: I'd get my money's worth, just like I used to pay for Qt. Sun had entirely too much overhead to let Java flourish. Java can support a small company: say 100-200 people, focused only on one product. Something the size of Sun? Not so much. Java could be monetized, just not by a big corporation. There is room for small, focused businesses doing what they do best. I somewhat dislike that Nokia took over Trolltech: sure the original owners/investors got their money's worth, but IMHO Qt lost out on it.

Re:Suicide? The end of java.

[parlancex]

What Microsoft actually pays for is licenses for patents pertaining to the similarities between the .NET Framework libraries and the Java standard libraries. I don't know the specific details but they are similar in that they both comprise a hierarchical set of namespaces and classes to provide certain functionality, and Sun had some patents applicable to that (apparently).

Re:Suicide? The end of java.

[ADRA]

I have little doubt that there are Sun (now Oracle) held parents that overlap with the offerings within CLI. The OP referred to James Gosling who would know surely know the truth of the matter given how close he was to Java within Sun for decades.

Re:Suicide? The end of java.

[KingMotley]

I think the OP was referring to a small interview done back in September in which Gosling was talking about Oracle, Sun, and Microsoft, and he said that the .NET framework runtime violated a bunch of Sun patents, not C#.

Re:Suicide? The end of java.

[kthreadd]

Objective-C is not a 'corporate' language. Some frameworks written in it is proprietary but the language itself is very much open.

Re:Suicide? The end of java.

[Pinhedd]

C# as a language might be standardized but there are potential problems with the run times themselves. Current .NET distributions contain both ECMA standardized libraries (up to 2.0 I believe) and those which are still proprietary (3.0, non ECMA and beyond, including WPF, WWF, ADO.NET, ASP.NET, Windows Forms, etc...). Some alternative CIL implementations have implemented more than the standard libraries and thus might run afoul of patents. Microsoft has been playing rather nice on the whole thing so far, even promising not to assert patents on many non-ECMA components.

Re:Suicide? The end of java.

[Simon (S2)]

Note that Microsoft heavily licences java to use in C#

[citation needed]

Re:Suicide? The end of java.

[KingMotley]

Unless I am mistaken, then 2003 revision included C# 2.0, and the 2006 version I linked to includes C# 3.0. Yes, WPF as far as I know is not. But then again, WPF/WinForms is a presentation/windowing library and can use QT or any of the other open source window libraries with it and skip WPF/WinForms if that really is your concern. Missing ADO.NET and ASP.NET are pretty big, but mono implements them anyhow. MySQL has it's ADO.NET provider that should work just fine on any CIL implementation as well.

And of course, you are more than welcome to implement any provider you so choose as well. The System.IO library that is standard in ECMA should be adequate to allow anyone to write whatever database driver necessary, or to implement an ASP.NET alternative.

Re:Suicide? The end of java.

[kthreadd]

So there needs to be a specification for a language to be open? Guess Perl haven't been open until Perl 6 then.

Apple have their GCC based compiler. GCC have their compiler. None of them is "more" Objective-C, just two different implementations.

Re:Suicide? The end of java.

[gbjbaanb]

nope. wrong. sorry. un-curl the corner of the C# carpet and you'll see Java.

from http://www.basementcoders.com/transcripts/James_Gosling_Transcript.html
Gosling's interview concerning Oracle's takeover of Sun.

James Gosling: ...I'm sure they were looking at the license fees they were getting from Microsoft. Microsoft .NET just smears over a huge pile of Sun patents. When they did the .NET design, they basically cut and pasted from the Java spec. The way that they did CLR, you know they swizzled the way the instruction set went but the way this thing really operated, they exercised essentially no creativity when coming up with .NET. They've done some things since then that have been kind of good but as part of the various court cases we ended up with this rather odd patent deal with them that involved them paying us fairly tasty amounts of money. And I'm sure that the lawyers looked at the Microsoft numbers and said, yeah I want that from Google

Your ISO standard is C# 2.0, not the current version - or did you expect better from Microsoft? I guess you did, more fool you.

Re:Suicide? The end of java.

[gbjbaanb]

see my reply to the other guy - from a /. story

Re:Suicide? The end of java.

[KingMotley]

Thanks for the quote that proves my point, it was regarding the .NET framework, not C#.

Re:Suicide? The end of java.

[shutdown -p now]

So there needs to be a specification for a language to be open?

He didn't say Obj-C is not open. He said that it is a "corporate language". Which means that this is a language which is primarily designed and implemented by a single company for the purposes of supporting their platform, and any third-party users or implementers - if they exist - do not have any real input on future language directions, and third-party implementations are marginalized. They can still be "open", it's just that no-one cares.

Both Objective-C and C# fit this really well. Java less so, but it looks like Oracle is determined to make it a part of their stack and nothing else.

Re:Suicide? The end of java.

[shutdown -p now]

You kinda miss the point of the whole "developers! developers! developers!" thing.

The reason why this has (and remains) a core part of MS strategy is because developers using MS technologies write apps which either work only on MS platforms, or work better on them than on other platforms. And app availability is what gets users onto your platform.

That is why, if anything, you see more free or very cheap development tooling (VS Express, SQL Express, DreamSpark & BizSpark etc) as competition ramps up. I fully expect that trend to continue.

Ultimately, all you need to do is look at the numbers. Windows and Office divisions are what brings in most of the cash in MS. Development tools are in supporting role there.

Re:Suicide? The end of java.

[shutdown -p now]

Unless I am mistaken, then 2003 revision included C# 2.0, and the 2006 version I linked to includes C# 3.0.

You're mistaken. The 2003 revision was for C# 1.1, and the 2006 one was for C# 2.0. This basically corresponds to .NET & Visual Studio release cycle - VS2003 / .NET 1.1 had C# 1.1, and VS2005 / .NET 2.0 had C# 2.0.

There has been no spec which covers C# 3.0 and above yet, even though the current stable release is 4.0. There is a published language spec directly from MS, but it's not Ecma or ISO standardized.

Re:Suicide? The end of java.

[CodeBuster]

Considering that C# is Java with a fair few mods.

I think you may be underestimating what C# brings to the table. The C# language has been very promiscuous over the years, borrowing the best and most clever syntax features from functional languages while improving the C/C++ style syntax to further favor the sort of "terseness" found in C/C++ but less so in Java. An opinion shared by many developers in recent years is that the Java language (i.e. the language itself, not as much the libraries) has fallen behind what C# now offers. Perhaps Oracle will work to remedy this situation? Perhaps, but for now if I had a choice between Java or C# I would choose C#. The following article is informative.

Re:Suicide? The end of java.

[Eskarel]

Sun lost for a number of reasons, not least of which was continuing to charge SPARC prices for standard x86 servers.

As far as Java went their biggest issue was the fact that while they came up with all the standards, their implementations sucked, and so no one bought their servers. Even their JVM implementation was worse than some of their competition. They tried to do something about this by following the same model they used with OpenOffice, trying to take advantage of free development from the open source community while selling premium supported editions, and Glassfish is actually now probably one of the best complete JEE containers available, but it was just too late.

Doing what they did was a huge gamble, it was always either going to completely destroy them or make them successful beyond their wildest dreams, unfortunately for SUN, the GFC happened right in the middle of their grand experiment and wiped them off the face of the earth.

The biggest issue with your statement however is that Oracle doesn't actually need Java to make any money at all, what they need is for Java to remain a serious contender to .NET. They need this for all sorts of reasons, but mostly because despite all their middle ware and acquisitions, their Database is still their bread and butter, and SQL Server is getting awfully close to parity for most use cases. Java is the glue that holds their stack together, and its survival is vital to their success.

That said, there isn't necessarily anything wrong with this particular move. The JVM is one of Java's major weaknesses at the moment, and if Oracle supports OpenJDK and gets a decent product out there, and then creates an enterprise ready version for a reasonable charge, Java could get better, not worse.

Re:Suicide? The end of java.

[KingMotley]

You are right. C# 3.0 spec has also been published here http://download.microsoft.com/download/3/8/8/388e7205-bc10-4226-b2a8-75351c669b09/CSharp%20Language%20Specification.doc, but it wasn't ECMA certified yet.

Re:Suicide? The end of java.

[whiteboy86]

Exactly, that was my point, thanks for elaborating.. I also like your bottom line motto.

Re:Suicide? The end of java.

[mcsandberg]

Objective C, while mostly used by Apple, is certainly not owned by them. Both it and C# are now part of the gcc, so arguably, neither is a "corporate" language. I actually use Objective C++ for OS X programming, darned if I'm going to give up Boost and the STL!

Re:Suicide? The end of java.

[mario_grgic]

Yeah, Objective C is more tied to single corporation than the other two. Apple totally and completely own the language, its runtime, libraries, and solely set the future direction for the language. Yes, they happen to use GNU toolchain, and contribute the objective C implementation to it, or rather they have their own private branch of GNU toolchain where they implement objective C language, but they are slowly moving towards clang/LLVM.

This is why you will not see the language used by anyone else other than Apple for anything interesting/important.

I don't get it

[rumith]

Either Larry Ellison is smart beyond my imagination, or he's too stupid to understand that he's basically killing MySQL, OpenOffice and Java - arguably the three most valuable software assets he bought with Sun.

Re:I don't get it

[Angostura]

I suspect that Ellison evaluates 'valuable' in straightforward monetary terms. "Is Java making me money? No? It's not valuable"

Re:I don't get it

[Sc4Freak]

Well, as a manager, that's his job - to maximise the wealth of the company shareholders. If Java isn't making money (directly or indirectly), then he needs to institute change so that it *does* make money one way or another. Otherwise he wouldn't be doing his job (and then he'd be fired and replaced by the board of directors).

Maybe tiering the JVM is the best way to maximise shareholder wealth, maybe it isn't. Either way, it should be obvious that any corporate manager would see "valuable" in a purely monetary sense. Even if Oracle decided to keep the JVM completely free, it'd be because they believe that keeping it free would make them more money than charging for it (maybe they think it'd drive sales of their other products or something like that, who knows). Everything a company does, even if it seems altruistic, is somehow contributing to the company's bottom line.

Re:I don't get it

[Anonymous Coward]

Maybe killing those technologies is something he wants... . It is not uncommon for companies to acquire another just to silently kill of the technology.

Personally I started looking into postgresql because to be honest I don't trust oracle and I'm certain mysql is the next thing onto the chopping block. It is a shame that a lot of providers (vps is to expensive here) don't see this coming and the number of hosts with postgresql support is very small.

Re:I don't get it

[Anonymous Coward]

There was always fight within SUN whether/how to monetize Java and JEE. Some wanted everything for free, hoping HW would sell and pay for SW; others wanted to go non-free way to charge for everything (Solaris). Some people wanted to use tiered approach (JEE/MySQL), giving the base for free and charge for special functionality/tools on top of that. The problem was that this strategy was changing every week. And then SUN was forced by investment funds to sell itself and the rest is a history...

Re:I don't get it

[drerwk]

...MySQL, OpenOffice and Java - arguably the three most valuable software assets he bought with Sun.

But not valuable enough to keep Sun in business for themselves.

Re:I don't get it

[Nursie]

"Well, as a manager, that's his job - to maximise the wealth of the company shareholders. If Java isn't making money (directly or indirectly), then he needs to institute change so that it *does* make money one way or another."

No, he has no such responsibility. He could decide that Java cannot easily be made profitable or sufficiently profitable and can or sell off the whole unit.

Re:I don't get it

[Savage-Rabbit]

Well, as a manager, that's his job - to maximise the wealth of the company shareholders. If Java isn't making money (directly or indirectly), then he needs to institute change so that it *does* make money one way or another. Otherwise he wouldn't be doing his job (and then he'd be fired and replaced by the board of directors).

I don't think Java it self is necessarily what you should look at to make money. It's products built with Java that make money. Products like Java stand and fall with a strong developer community and convincing a significat portion of the worlds developers to use a some programming language is very, very, very hard. Universities, for example, are full of people who tried and failed to create and popularize new programming languages and the same goes for all sorts of companies that had the same idea. Making all of Java available for free was the best way to assure that it was widely adopted by a strong developer base. Had Sun made any serious attempt to charge for a "premium" edition or perhaps giving away the "basic" Java package and selling "advanced" features it would have caused Java to fail. You are better off making only a "premium" edition and distributing it for free because otherwise any developer using the "regular" edition gets the feeling he/she is a second class citizen, never mind the fact that startups for example frequently need the "advanced" or "premium" features but can rarely afford to pay insanely expensive license fees to pay for them so they will typically choose something free and that adversely affects the growth of your developer community..

Re:I don't get it

[Savage-Rabbit]

e could decide that Java cannot easily be made profitable or sufficiently profitable and can or sell off the whole unit.

Oracle sell off the Java and MySQL units to somebody who knows what they are doing? Hope springs eternal...

Re:I don't get it

[Nursie]

Pretty sure that poster was saying that Ellison had a duty to 'save' Java, which is what I was objecting to.

Re:I don't get it

[EsbenMoseHansen]

Postgresql is superior to mysql in nearly every way imaginable, so that would make sense in any case. It can also distinguish between null and the empty string ;)

Re:I don't get it

[Kjella]

No, but you have companies that take the physical assets like office chairs and tables and liquidates those. Unless Oracle wants to go into that line of business, they ought to have some good plans for what to do with the business assets. I get it that some things you cut and some you keep and those you cut you monetize. Like if you bought a farm with corn and livestock but want to focus on corn, you send the livestock to the slaughterhouse. But so far I've seen nothing but slaughter in the software assets, and I really can't believe the Solaris hardware/OS is that valuable. They got serious high-end mainframe competition by IBM mainframes with DB2, strong hardware competition from Intel and AMD and strong software competition from Linux. Granted, they did stop MySQL but on the whole that's a rather expensive price to pay.

Re:I don't get it

[gilesjuk]

Oh I think he is very clever. He has as much wealth as Bill Gates and others. That's not by accident.

MySQL is pitched as their entry level database offering. There are many MySQL experts who won't want to spend time learning another database engine. MySQL is a brand, there are alternative forks of it, MariaDB. But they don't have the brand name which is familar.

Java is used in lots of large web applications. There has always been people who wanted more performance or some commercial help. They will now be able to pay for that extra performance.

Re:I don't get it

[andre1s]

Hmm given a task at hand e.g. making $ I would have to guess Larry "is smart beyond my imagination" applies here given he is one of world's 10 richest people.

Re:I don't get it

[Haedrian]

You actually think that they bought Sun for their software?

Its all about the patents. So they can sue larger companies for making Android.

Re:I don't get it

[fatp]

MySQL: AFAIK, the breakup is ignited by Monty (the original developer), who sold the product to Sun for 1 billion, demanding Oracle to give everything for free. All these before Oracle was able to do bad anything to the project. And I think PostgreSQL is a better FOSS RDBMS.
OpenOffice: Actually I don't know what bad thing had Oracle done to cause the breakup. But surely the developer know it better. I think this can be a sample case on how Oracle treats FOSS projects.
Java: As others said, Oracle has always been selling Jrockit. So making a for-free and a for-fee JVM seems acceptable... as long as they are compatible.

Actually, I would kick MySQL from the "most valuable software assets", and add (Open)Solaris and VirtualBox to the list...
(Open)Solaris: This is obviously an evil thing Oracle doing to the FOSS community. But I am not sure whether it's reasonable to extrapolate to other projects.
VirtualBox: This seems something that's impossible to draw Oracle's interest (Oracle sells XEN based Oracle VM). It is really surprising that it still has quite regular releases...

Declaration of Conflict of Interest: I am an Oracle DBA

Re:I don't get it

[drerwk]

There are a whole lot more people that second guessing Larry Ellison than providing an equivalent value to their shareholders. Not that I have experience in billion dollar acquisitions, but the 50 and 100 million dollar acquisitions I've seen first hand took years to play out.

Re:I don't get it

[Anonymous Coward]

Anybody who can take the steaming pile of monkey dung of a db that Oracle is and make yacht-loads of money from it, is truly a genius.

Even with as much as I can't stand Oracle, I have to admire Ellison and their leadership for their unbelievable business capabilities.

Re:I don't get it

[obarthelemy]

how are those assets valuable to him ?

what genius plans would you implement for them instead of his ?

Re:I don't get it

[mfnickster]

Oh I think he is very clever. He has as much wealth as Bill Gates and others. That's not by accident.

Actually, he has about half as much as Bill Gates, but your point still stands. He knows what he's doing.

http://www.forbes.com/wealth/forbes-400/list

Re:I don't get it

[Steauengeglase]

Will, never, ever, ever, ever happen. Java and MySQL are Oracle IPs now. Even if they are completely worthless on their own, why give up the chance to sue someone? All I can say at this point is fuck 'em and good luck to all of those forked projects.

Re:I don't get it

[Anne Thwacks]

They got serious high-end mainframe competition by IBM mainframes with DB2, strong hardware competition from Intel and AMD and strong software competition from Linux.

Yes, those of us that do financial transactiopns on postgresql/OpenBSD/sparc64 are hard at work planning a migration path, cos remaining with sparc64 looks very risky, and most of us have already tried Oracle, thats why we are not using it now.

Re:I don't get it

[Anne Thwacks]

Mod parent +infinity: he's got it!

Re:I don't get it

[SanityInAnarchy]

I'm not sure when that's a feature. It appeals to my inner software developer, but at the same time, Ruby on Rails has this feature I use all the time -- the ability to deliberately not distinguish between null, an empty string, and a string containing only whitespace.

nil.blank?
''.blank?
' '.blank?

all return true. This more or less matches my experience with Postgres vs MySQL -- Postgres made more sense in many ways, but MySQL had more stuff that just worked, and in the end, neither of them were quite as sexy as CouchDB and friends.

Re:I don't get it

[turkeyfish]

Would be curious as to what you would see as a better alternative? C++? .Net? C#?

The first could be a killer, but so far puts the onus of memory management on the individual developer and hence leading to highly inconsistent even if extraordinarily capable apps across systems. .Net and C# are highly single platform centric. If you are tied into the right platform, you may be in good shape, but not if you seek cross platform compatibility, especially important for ubiquitous web development.

Its hard enough for a single developer to learn even one of these large systems well, much less try to get them all singing in tune across all platforms. Ultimately, all platforms lie or die by the number of developers they are adding to their universe.

Re:I don't get it

[No. 24601]

Either Larry Ellison is smart beyond my imagination, or he's too stupid to understand that he's basically killing MySQL, OpenOffice and Java - arguably the three most valuable software assets he bought with Sun.

It's easily argued that these three most valuable assets acquired in the Sun purchase are, taken as a whole, not even worth a fraction of the existing and potential business of the product they've already had, Oracle. They couldn't give a shit about Java, MySQL or gawd OpenOffice. It's more likely he was thinking: "hmmm, how do I squash three great products at least one of which had terrifying potential???"

Re:I don't get it

[CrashandDie]

So you think Oracle bought Sun because it was a company that was losing money?

Or because it provided Oracle with opportunities to make money?

Re:I don't get it

[hairyfeet]

I don't think so. Old Larry has talked for years about the "full stack approach" done by IBM, and by buying SUN he has that now. Here is what I think will happen:

Oracle will play lip service to their Linux, while slowly letting it die. FOSS doesn't play into their full stack so frankly they don't care if the community pisses up a rope. What they will do is have a top to bottom "one chicken to choke" Oracle DB solution, with custom SUN hardware running a customized Solaris, and both will have been tweaked for maximum throughput for Oracle DB.

Now considering how many enterprises are hooked on Oracle DB the way desktop users are hooked on windows? This approach will be VERY profitable, and make old Larry another couple of mountains of money. As for Java? I hate to tell you FOSS guys but Larry doesn't care about you because you are NOT his customers nor his target audience. What IS of concern to old Larry is there is a hell of a lot of Java code made by many (including Oracle) to interact with Oracle DB. And THAT is what will most likely get the "premium" experience.

From his past actions old Larry couldn't care less if FOSS guys get pissed or not he is there to make the big bux and from his track record he will make a shitload off of this, whether FOSS guys like it or not. Lucky for all there are choices out there, the question is if the expense of migration is worth it.

Re:I don't get it

[Raenex]

The point is it won't be "business as usual" under Oracle. Sun was barely staying afloat and desperately looking for a buyer.

Re:I don't get it

[dieth]

Ya but you'd have to euthanize all those poor fools that can only write in Java and can't wrap there minds around any other language. (99% of the 'Java' coders I've met throw a shit fit if they're asked to write anything in C)

Re:I don't get it

[jez9999]

Either Larry Ellison is smart beyond my imagination, or he's too stupid to understand that he's basically killing MySQL, OpenOffice and Java

3 opensource projects. Someone please tell me what I'm missing here. How can one company 'kill' an opensource project?

Re:I don't get it

[Trailer Trash]

I've mulled this over in my head a lot. The guy's worth tens of billions of dollars, so I'm not going to sit here and question his judgement, call him an idiot, or any of that. At the same time, I believe that his attitude - which seems to trickle throughout his company - has a negative impact on his company's performance.

But in the end, does it matter for him? What if he could make a few billion more off Java in another way? Does he need the cash?

Re:I don't get it

[rtfa-troll]

Would be curious as to what you would see as a better alternative

How about: native code so it can be fast and light weight + a variety of languages so we can choose the right one for the job + some kind of application security system (SELinux?) + use of source code so that it can be portable + use of QT so you can have a portable graphics library?

Java was meant to be build once run anywhere, but it's turned out not even able to run in one place since you need multiple JVM versions for different applications. Once that promise goes what do we have? A pretty efficient byte code interpreter with a wide set of libraries. Some good security mechanisms and quite a bit of holes.

You should note, that once you are in a QT environment, I believe that even if you aren't doing Graphics programming, almost all of the self allocation is gone, which puts you pretty much in the same state as .Net. You can do it if you want to but it's not done by default.

Re:I don't get it

[rtfa-troll]

Alternatively just have them compile their java source to native code and integrate them into a normal environment. GCJ should be a good place to start.

Re:I don't get it

[drerwk]

I think Oracle bought Sun because Sun was cheap to buy, and Oracle had a plan on converting the purchase into more money than they spent. Why was Sun cheap? Because the market had lost any expectation that the current management could earn any value. That is why the Sun stock was down 80% over the four years prior to the sale.

Re:I don't get it

[masterwit]

Didn't you see the tag on the article
(idiots I believe)
So to comment on what you said, I think that Larry Ellison is one of the many idiots over there: we should not be so hasty to just give one guy at Oracle all the credit...although he is the emperor in that death star they call Oracle. :)

Re:I don't get it

[M. Baranczak]

If the JVM is patent-encumbered, then it's not really open, so Larry can do whatever he wants with it. I'll be watching the Google lawsuit pretty closely.

Re:I don't get it

[Tim C]

"Is Java making me money?"

Yes, it is.

Pay for performance?

[whiteboy86]

If you want performance then you have better options, some of the IEEE standard languages for example and that is what pros like Ebay or Google are using anyway, not Java. Plus you buy yourself some freedom from the corporate control like this.

Re:Pay for performance?

[andre1s]

You do realize that Google uses Java extensively ?

Re:Pay for performance?

[Anonymous Coward]

IEEE standard languages

C and C++ are ISO standards. What high performance languages are you thinking of?

Re:Pay for performance?

[Xest]

...and rather amusingly, eBay runs on Java too.

In fact, he chose about the most prominent two companies in the IT world that do use Java. Microsoft, Facebook, Twitter, and I believe even Amazon, do not use Java.

Re:Pay for performance?

[mcvos]

Exactly. Google uses so much Java I'm surprised they weren't the ones to buy Sun. How much better the world would have looked if they had.

Oracle is doing everything they can to fuck up

[Sycraft-fu]

I'm going to laugh as their Sun acquisition goes down in flames and they end up losing money on the whole deal. They seem to be working to identify any market they can that things are working in and eliminating it. They've done a great job at getting us to work at getting rid of all our Solaris systems as fast as we can.

While in theory this could be fine for Java, I can't imagine it will be being how poorly Oracle has handled things so far. Most likely it'll be a case where the free JVM will be a piece of crap on purpose, and the pay for JVM will be required for anything to work well. Ya, well, that'll fly like not at all. People are not going to go and buy something to make Java apps work better. Perhaps companies who rely heavily on Java on the back end will, but more likely they'll just stop upgrading and switch to something else.

I guess we'll see, maybe I'm wrong and the premium version of the JVM really will provide worthwhile premium features that high end users want, while the normal JVM remains for normal people. However I doubt it. I think they'll try and charge every person for the JVM on their computer, which just won't fly.

Re:Oracle is doing everything they can to fuck up

[polemon]

As I understand it, when the open source version gets adopted by a voluntarily group of individuals that keep developing the open source version, it might be forked pretty soon, so nothing of value would be lost.

As for the commercial version, that is probably gonna end like many other Oracle products, that got forgotten. I'll linger around in Oracle's inventory, but nobody will care much about it. Also, the fork might overshadow the commercial version in a couple of months, since performance tunes, are not exclusive to Oracle programmers...

Re:Oracle is doing everything they can to fuck up

[jonwil]

If Oracle starts locking things up in the premium version, OpenJDK will be forked (there are already some shallow forks like IceTea that take OpenJDK and replace the remaining closed-source bits with stuff from GNU Classpath etc) and the community will shift.

Its happened to OpenSolaris with the Illumos project and OpenOffice with the LibreOffice project.
No reason it cant happen with OpenJDK.

Although what might happen is that Oracle will find a way to write various APIs and licenses such that if you copy certain features from "Java Premium" you loose the patent grant given under the OpenJDK APIs.

Re:Oracle is doing everything they can to fuck up

[sg_oneill]

Oh, the dot net crowd are going to eat Oracles lunch over this. Microsofts dot net is free as the wind (well not free as in speech, but whatever) and has a lot of "enterprise-y" features.

Am I the only one seeing echos of the sad demise of Borland into irrelevance here?

Re:Oracle is doing everything they can to fuck up

[DrXym]

People are not going to go and buy something to make Java apps work better. Perhaps companies who rely heavily on Java on the back end will, but more likely they'll just stop upgrading and switch to something else.

I'm surprised Oracle even have such a grip on Java as they have. Why doesn't someone produce a comprehensive open source test suite analogous to the real certficiation tests? Then who cares if a JVM is officially blessed Java or just some offshoot.

Re:Oracle is doing everything they can to fuck up

[ChunderDownunder]

There is one for Classpath called Mauve. I doubt it's 100% comprehensive but it's a start.

Re:Oracle is doing everything they can to fuck up

[shristov]

According to Oracle (TFA), "There will always be a high-performance gratis JVM." Well, perhaps Oracle are just going to add enterprise-class features to Java - for example, the JRockit hot swapability mentioned in the article. Once you need such features, chances are you are able and willing to pay for these. The rest of the community could continue using Java for non-mission critical purposes. In time we'll see if this strategy is successful, or not. If demand for features like the ones Oracle is planning to develop is great enough surely open alternatives for some of these will pop up in foreseeable future. When/if this happens we'll hit the major issue worth discussing: how the Oracle-led and OpenJDK evolution paths will stay at least close to each other. If they diverge substantially, at least one would be doomed... if not both.

Re:Oracle is doing everything they can to fuck up

[ultranova]

If Oracle starts locking things up in the premium version, OpenJDK will be forked (there are already some shallow forks like IceTea that take OpenJDK and replace the remaining closed-source bits with stuff from GNU Classpath etc) and the community will shift.

I'm sure that Oracle will find some obscure patent- or other issue to crush the free version. That is what patent law exists for, after all: to help build monopolies.

Oh well, I guess it's time to start looking for another language to start new projects in.

Re:Oracle is doing everything they can to fuck up

[ikkonoishi]

I'm guessing it will be something like
iBWAHAHA = 0
do {
wait(10);
iBWAHAHA++;
} while (iBWAHAHA != 10|bPaidOurExtortion);

Re:Oracle is doing everything they can to fuck up

[shentino]

It could be that Oracle smelled blood and decided to go for the kill.

Interesting that soon after acquiring them, they sue Google over patents.

Re:Oracle is doing everything they can to fuck up

[ubersoldat2k7]

COBOL?

Re:Oracle is doing everything they can to fuck up

[Lennie]

"it might be forked pretty soon, so nothing of value would be lost."

compatibility ?

Re:Oracle is doing everything they can to fuck up

[AmberBlackCat]

I really don't think they can make it much better than it is. So instead, they're going to take all of the minor bug fixes and optimizations that used to be in the free version, and only put them in the expensive version, leaving the free version to never see a real update again. It's not about making the expensive version better. It's about making the free version worse.

Re:Oracle is doing everything they can to fuck up

[Lennie]

That is not interresting, that is to be expected. Because it has already come out that they explicitly wanted to know about the patents and Google, it was even suggested it was a dealbreaker if it wasn't something they could sue over.

Re:Oracle is doing everything they can to fuck up

[evilbessie]

They'll just make it compulsory to access their databases, at least all the Oracle stuff where I work requires Java. It might convince shops to give up Oracle, or being as they have already bought into it, they might just keep paying...

Re:Oracle is doing everything they can to fuck up

[Col Bat Guano]

"They've done a great job at getting us to work at getting rid of all our Solaris systems as fast as we can."

That's ok, you can always switch to Mac OS X and XServes... oh, wait....

Re:Oracle is doing everything they can to fuck up

[dna_(c)(tm)(r)]

Once you're convinced you need such features, chances are you are able and willing to pay for these. The rest of the community could continue steering clear of using vendor lock-in bait for their mission critical applications in Java. It is the classical trick of offering a small incentive so that your codebase is no longer vendor neutral.

Most Application Servers support reloading classloaders, so you can already restart apps without restarting the AS (or the OS of course).

If we're talking about hot swapping parts of code (bug fixes etc. like you can in a debugger) then this will introduce a whole new class of problems - what version of the software is running at a particular point in time? Kind of Continuous Disintegration...

Re:Oracle is doing everything they can to fuck up

[Teckla]

While in theory this could be fine for Java, I can't imagine it will be being how poorly Oracle has handled things so far.

From my perspective, it's perfectly fine if Oracle has decided to try to aggressively monetize Java. The real problem, in my opinion, is their lack of clear and detailed communication.

To announce there will be two JVMs without giving us details is insanely stupid of them. It leaves developers, like me, uncomfortable with moving ahead on Java-based projects. It's the not knowing that's the killer. Will the free version continue to meet our needs? We don't know, because Oracle hasn't given us any damned details. Just some vague announcement that's leaving everyone uncomfortable.

The same applies to Java on OS X, too. Oracle, once again, leaves us wondering what they'll do. They should have already announced their intention to either pick up where Apple is leaving off, and ship future versions of Java for OS X directly, or not. That way, the open source community could make a decision whether or not they want to do that work. And then developers that want to also target OS X could start making some decisions.

But no, Oracle is being tight lipped, leaving OS X folks wondering and uncomfortable about the future of Java on OS X.

Oracle just sucks at communication, and I've already halted my personal Java projects, and have started seriously considering alternative technologies to replace Java.

Re:Oracle is doing everything they can to fuck up

[modmans2ndcoming]

Mono is pretty much a drop in replacement for Microsoft's platform. I am sure it will become much more popular when Java dies.

Re:Oracle is doing everything they can to fuck up

[modmans2ndcoming]

What Oracle should do is put their development time behind build an good high performance and non-buggy Java app server. Websphere and JBoss suck. Most companies would pay handsomely for an App server that makes life easier and gives their end users better performance and reliability.

Re:Oracle is doing everything they can to fuck up

[turbidostato]

"Yes, Borland is the first that comes in mind. Well, it's time to start a new company based on ideals that permeated Borland, SUN, SGI and all other pioneers of open standards."

That's all well and good, you know, but... how do you expect to monetize those ideals?

I try to imagine somebody going to a VC firm saying he's gonna try to retake the ideals that permeated Borland, Sun and SGI. I think the first the VC people would say is "What? those companies that broke? Why do you think taking visions that broke those companies is such a good idea?"

Re:Oracle is doing everything they can to fuck up

[PastaLover]

No reason it cant happen with OpenJDK.

AIUI there is a major problem, one that apache Harmony is now faced with. Basically, to get a patent grant for your open source project you need to show that you have delivered a full implementation of the platform. To show that your implementation needs to pass the TCK tests. To get those tests (that are proprietary software, owned by Oracle), you need to agree to certain Field of Use restrictions. Which are incompatible with pretty much any open source license you can name.

So while OpenJDK has Oracle's blessing and thus gets to get out from under this problem, any other open source project that forks off OpenJDK would lose access to the TCK and probably find itself in Oracle's crosshairs a couple months/years down the line.

Re:Oracle is doing everything they can to fuck up

[PastaLover]

If you don't pass the official testsuite, you don't get the patent grant. So Oracle can simply sue you.

Re:Oracle is doing everything they can to fuck up

[binarylarry]

Mono isn't .NET, its an incompatible reimplementation of the platform. It's missing many, many things. That said it's fairly decent given the source.

OpenJDK on the other hand *is* the same codebase as Java. It is Java... except it's open source. HotSpot, the vm that's the heart of the Java runtime, is included and it's much faster and more reliable than Mono.

Re:Oracle is doing everything they can to fuck up

[Haeleth]

"Guys, Java is dead because it was bought by a litigious patent-loving company with monopolistic ambitions and a history of screwing its customers."

"Fear not! We'll just switch to a semi-compatible clone of a semi-closed platform owned by a different litigious patent-loving company with an actual monopoly and an even longer history of screwing its customers. Problem solved!"

Re:Oracle is doing everything they can to fuck up

[Bj�rn]

Oracle already bought the BEA WebLogic server. That is where the JRockit VM comes from. And when they acquired Sun they got the Glassfish App sever. Glassfish is one of the few Java EE 6 certified severs.

I'm guessing and hopping that the intended market for the premium VM will just simply be for WebLogic customers and the like. In other word customer, like banks, who have money to spend om important server applications.

Re:Oracle is doing everything they can to fuck up

[gtall]

I don't think Oracle's intent of the lawsuit is to monetize Java, That''s merely a collateral benefit. Rather, Google does present a threat to Oracle. Oracle sells data retrieval. Google's business is built on data retrieval. Given a few years of no real competition, Google will start threatening Oracle's schtick. Uncle Larry is too paranoid not to see Google as a threat, so he looks in his bag of tricks to see how to knacker them now.

Re:Oracle is doing everything they can to fuck up

[makomk]

The same applies to Java on OS X, too. Oracle, once again, leaves us wondering what they'll do. They should have already announced their intention to either pick up where Apple is leaving off, and ship future versions of Java for OS X directly, or not.

That assumes that Oracle know what they're going to do. I would be entirely unsurprised if they only found out about this at the same time as the rest of us - or possibly even a bit later - and haven't managed to come to a decision yet. It seems to be Apple's usual MO.

Re:Oracle is doing everything they can to fuck up

[NoOneInParticular]

Companies that power their business (website + internal apps) using application servers, will easily pay up for an optimized JVM. They already pay for the hardware, the operating system, the database, the application server, so why not for the JVM? As JVMs are now monetizable, another fight will break out between Red and Blue (Oracle and IBM), this time over the speed of their JVM. IBM will start to pour more money in their JVM research, especially when Oracle is going to make running with WebSphere less comfortable.

Of course, Java for consumer or startup level dies in the process, and the inevitable future of Java as the new Cobol will become a fact. Oracle and IBM will milk this cow forever, while the rest of the world moves on.

Re:Oracle is doing everything they can to fuck up

[JAK]

Microsoft .Net is free? Really? The cheapest version of visual studio appears to be $800 (and that version has far fewer "enterprise-y" features than the $5500 premium or $12,000 ultimate versions). I realize there are "free" parts of .Net, including the express editions of visual studio but there are already arguably superior equivalents available in the Java community.

Re:Oracle is doing everything they can to fuck up

[TheTrueScotsman]

The language doesn't matter; the runtime does. If lambdas or functional transforms or dynamic objects matters to you, you will already have implemented it in some cool libraries and got on with shipping stuff (my 20% is my weekends...).

All I can say is That JVM has excellent performance and compatibility for the OSes I use on the server side (and indeed the Java language has all the primitives I need); if Oracle can give me something significantly better, I'll be prepared to pay for it.

Re:Oracle is doing everything they can to fuck up

[Necroman]

I wish I had some points to mod you up with.

The uncertainty is definitely frustrating. They are trying to make a lot of changes to java (this being the latest) but we aren't being given all the facts.

Thing is, I don't think it's time to jump ship just yet. Oracle hasn't done anything terribly stupid yet, they are just setting themselves up to do something stupid. It's possible they will see the public nervousness and thrown us a bone... but who knows.

Re:Oracle is doing everything they can to fuck up

[tokul]

Oh, the dot net crowd are going to eat Oracles lunch over this. Microsofts dot net is free as the wind (well not free as in speech, but whatever) and has a lot of "enterprise-y" features.

Go easier with stuff you are smoking. You missed dependency on OS and price that OS in enterprise environment. .NET is not free. It is only freely distributed addon of commercial operating system.

Re:Oracle is doing everything they can to fuck up

[jack2000]

Your sales pitch is bad. Instead go: "gonna try to retake the ideals that permeated Microsoft, HP, [whatever big company]. And THEN when they buy it hook line and sinker you try and not be evil.

Which is hard apparently even for google.

Re:Oracle is doing everything they can to fuck up

[turkeyfish]

Thats going to be tough considering that more most corporations and large non-profits, changing software platforms is not as easy as just changing particular software applications. Johnathan Schwartz, the last pre-sale Sun CEO, pretty much opened up Java the language to open source prior to leaving and prior to the sale to Oracle. No way legally Oracle can get that worm put back in the box.

You might be right on the JVM side, but its not as if the open source community can not build an alternative JVM, especially if Google wins the lawsuit, which in my opinion is doubtful because the courts don't understand tech only the need to maintain a pro-corporate status quo. However, whether an open source JVM will benchmark favorably enough to the proprietary one offered by Oracle to be viable in the future remains to be seen.

These issues will sort themselves out in the next couple of years as we see what posture Oracle settles on. So far, their PR has just been terribly damaging to the overall "Java" brand. Greed does have a way of undermining the cleverest of strategies.

Re:Oracle is doing everything they can to fuck up

[Lost Engineer]

Royalty free.

I didn't RTFA and the summary doesn't make it clear, but it sounds like at least one possibility is that Oracle intends to charge for the premium JVM, which would need to run on each server.

Re:Oracle is doing everything they can to fuck up

[turkeyfish]

Only if you choose to call it "Java".

Oracle does not own the patent to the concept of a VM based language, only the patent to the JVM. A problem for Oracle is that Sun open-sourced much of the Java language prior to the sale, just not the JVM.

That concept of a VM based language originated with UCSD Pascal, long before Java ever appeared as a language. If Google can convince the courts that their implementation is not a JVM per se, only a VM, its on to the races and there will be for the first time, true competition that can't be easily monopolized.

Re:Oracle is doing everything they can to fuck up

[cecom]

If Oracle starts locking things up in the premium version, OpenJDK will be forked (there are already some shallow forks like IceTea that take OpenJDK and replace the remaining closed-source bits with stuff from GNU Classpath etc) and the community will shift.

Its happened to OpenSolaris with the Illumos project and OpenOffice with the LibreOffice project.
No reason it cant happen with OpenJDK.

Not going to happen. HotSpot is way too complex to be maintained by the "community". As far as I know, so far, years later, the "community" hasn't been able to port HotSpot to a single additional architecture. It runs in interpreted mode, but not as JIT on anything other than x86 or SPARC.

Re:Oracle is doing everything they can to fuck up

[Mongoose Disciple]

I realize there are "free" parts of .Net, including the express editions of visual studio but there are already arguably superior equivalents available in the Java community.

Ugh. Eclipse. Do not want.

Re:Oracle is doing everything they can to fuck up

[Mongoose Disciple]

Johnathan Schwartz, the last pre-sale Sun CEO, pretty much opened up Java the language to open source prior to leaving and prior to the sale to Oracle. No way legally Oracle can get that worm put back in the box.

I think (unfortunately) you're really underestimating the army of lawyers Oracle is willing to throw at this problem -- or maybe you're thinking that because, legally, Oracle may not have a leg to stand on they can't still get what they want.

Sue enough people for long enough -- even if you ultimately lose every case -- and some will figure it's easier/cheaper to just buy Oracle off; others will move to another platform.

Re:Oracle is doing everything they can to fuck up

[modmans2ndcoming]

MS is litigious?

Saying doesn't make it so.

Re:Oracle is doing everything they can to fuck up

[tokul]

ALL of the libraries, frameworks, languages on .net are free.

Not all. Windows is not free OS.

Re:Oracle is doing everything they can to fuck up

[PincushionMan]

That works well... right up until the time that Oracle sues the pants off of you. The problem is that Android is based on Apache's Harmony JVM. Which just about everybody other than Google distancing themselves from, most importantly IBM - they've been fans for 5 years now.

Java the language may be free, but Java the VM and Java the Classpaths are less so. I believe that Sun said they'd never act on their patent portfolio and harm the community. Oracle is the total opposite.

Re:Oracle is doing everything they can to fuck up

[PincushionMan]

No no no. COBOL.NET. What? It's not a joke? Dude!

Re:Oracle is doing everything they can to fuck up

[rtfa-troll]

• Microsoft sues TomTom over Linux and other patent claims
• Aiming at Android, Microsoft sues Motorola
• Microsoft sues Salesforce.com for alleged patent infringement

That's before we get to the actions of the major Microsoft shareholders e.g: Microsoft Co-Founder Launches Patent War "

And finally of course ; Microsoft's apparent involvement in many proxy actions.

• Microsoft Proxy Fights Against Google in the United States
• Microsoft Proxy Attack on GNU/Linux Continues With TurboHercules
• Google Accuses Microsoft of Proxy Legal War
• Also suggestions of MS involvement in the SCO lawsuit

Under previous management MS may not have been lawsuit happy. Nowadays they pretty clearly are.

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

If lambdas or functional transforms or dynamic objects matters to you, you will already have implemented it in some cool libraries

You can't implement closures (Lambdas) in a library - it's a fundamental language feature. It's either there or it's not.

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

try looking at the freaking link moron. Very little is missed in Mono.

Yeah, only all the new - as in, introduced in 2006 - APIs such as WPF (completely missing) and WCF (early alpha stage).

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

Only if you choose to call it "Java".

Name is a trademark. GP was talking about patents.

Oracle does not own the patent to the concept of a VM based language, only the patent to the JVM. A problem for Oracle is that Sun open-sourced much of the Java language prior to the sale, just not the JVM

Oracle does not own patent to JVM (you cannot patent specific software). But they do own patents to many VM implementation techniques that are, supposedly, very hard to avoid when writing an efficient VM - especially a Java VM. That is precisely what they used to sue Google over Dalvik, which doesn't even try to implement the JVM spec!

That concept of a VM based language originated with UCSD Pascal, long before Java ever appeared as a language.

Did UCSD Pascal have a JIT-compiler for its bytecode?

Re:Oracle is doing everything they can to fuck up

[Eskarel]

I'll agree on most things. Oracle has been incredibly tight lipped on all sorts of things since they started the Sun acquisition, much to the detriment of the things they have since acquired.

OSX though isn't really their fault. I'm fairly convinced that Apple pretty much just sprung that on them, and without knowing whether Apple will provide them with the source of what the current version or publish(even under an NDA) the APIs that it uses to actually access the Apple GUI, it's just not a decision they can make. Lord Steve is engaged in a war with anything which doesn't fit into his vision of what the world should be, and it's entirely likely that, for whatever insane reason, he actually wants to kill Java on Apple platforms, if I were an OSX developer I'd be a lot more concerned about what Lord Steve is planning than on what Oracle is going to do.

Re:Oracle is doing everything they can to fuck up

[exomondo]

Java the Classpaths

The Classpaths?

Re:Oracle is doing everything they can to fuck up

[TheTrueScotsman]

Nonsense. In Java just wrap it in a class. E.g (Let's call the class lambda and have it predefined with a single func method, say in a library):

List<Lambda> myListOfLambdas = new ArrayList<Lambda>();

public void letsMakeALambda(){
    final String localContext = "A proper closure sees the local context";
    myListOfLambdas.add(new Lambda(){public Object func(Object o){
        return localContext;
    }});
}

public void letsUseALambda()
    for(Lambda l: myListOfLambdas){
        System.out.println(l.func(null));
    }
}

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

Nonsense. In Java just wrap it in a class. E.g (Let's call the class lambda and have it predefined with a single func method, say in a library):

The only reason why you can do it is because Java anonymous inner classes are closures - they capture the local environment, even if not fully (no mutable locals). Now try doing the same in C++. Or in Java 1.0 (which didn't have anonymous inner classes).

Re:Oracle is doing everything they can to fuck up

[Ash-Fox]

try looking at the freaking link moron. Very little is missed in Mono.

That link doesn't really reveal how much isn't implemented. For example, none of my J# applications work in Mono, no matter what version of .net it's compiled for. The poor MSIL support isn't mentioned on that page, is it?

Re:Oracle is doing everything they can to fuck up

[magus_melchior]

No reason it cant happen with OpenJDK.

To add to PastaLover, Oracle is already doing what you're saying is impossible-- they're suing Google over Android.

Re:Oracle is doing everything they can to fuck up

[modmans2ndcoming]

who cares about WPF? GTK# is available to code in and provides complete cross platform abilities. WCF would be nice but is not required.

For someone starting a project, it is easy to build using Mono. For someone wanting to be cross platform using c#, mono is easy to use and it is very easy to avoid using WCF and WPF. WCF will be fully supported with in the next year.

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

who cares about WPF? GTK# is available to code in and provides complete cross platform abilities.

The person whom you told is a "moron" said that Mono is missing many things that are there in .NET, and is therefore not a "drop-in replacement" which you claim it to be. WPF is one such major thing. That Mono has other options to offer is irrelevant - if I have an existing .NET codebase, you can be certain that it won't use Gtk#; and if it's written in the last 2-3 years, it's quite likely to use WPF. I'll have to rewrite that to port it to Mono.

Then also, going with Gtk# means no designer support in VS, which is still the prime .NET development environment. Sure, one can use SharpDevelop on Windows - and miss all the features VS has over that (not to mention stability... for me at least, SharpDevelop has consistently being a crashfest).

Yes, you can use Mono as a cross-platform development environment if you knowingly and deliberately target it and account for its limitations. No, it's not a replacement for .NET.

Re:Oracle is doing everything they can to fuck up

[Ash-Fox]

Considering Mono supports C#, VB.Net and not J# would be reason for that.

That's compilation, I'm talking about supporting the MSIL, what every Microsoft .NET language is transformed to before converted into bytecode. A proper .net runtime/interpreter wouldn't have these problems, regardless. By the way, MSIL is a supported language of Mono, since it has to use it to begin with.

Besides... only a loser would use J#

The loser here is Mono and .NET.

People like you mention migrating to .NET from Java is better, but then it's attempted and cross platform support completely sucks in comparison to Java. Why are you trying to miss lead people into making obviously flawed decisions?

Now, getting back on topic.

The poor MSIL support isn't mentioned on that page, is it?

Re:Oracle is doing everything they can to fuck up

[TheTrueScotsman]

Well, we are talking about Java after all (and surely not 1.0 which only has historical interest). Mutable locals are easily achieved by putting them in a collection - the collection reference is final and immutable but its contents aren't.

Re:Oracle is doing everything they can to fuck up

[shutdown -p now]

Well, we are talking about Java after all

No, not really - we're talking about whether the choice of language matters or not, which can only be viewed in the context of all languages out there. My point, again, is that you cannot always substitute for language deficiencies with library solutions. In this particular example, you can use closures because Java has them. In a language which does not, you cannot do so.

You can hand-code a closure as a class, of course, with explicit capture - but this is not a true drop-in replacement, since it requires dropping to significantly lower levels of abstraction. You can also implement OOP as a library in C, and you can implement all C constructs in assembler - but it's not a viable replacement.

What's interesting is that language design, as of late, has been very much about actually allowing what you speak of - making the language itself extensible to the point where, if something new is needed, it could be implemented as a library. Closures and macros are where it's at. Give it 10-15 years, and we'll end up with Common Lisp with curly braces (since it'll be derived from Java and C#).

Re:Oracle is doing everything they can to fuck up

[PastaLover]

The problem being that if you implement Java or a Java-like language right now, you're not sure whether you won't be staring at the business end of a million dollar lawsuit in a year's time. So people will wait out what happens in the Google lawsuit, which might take years (actually, if it's like SCO at all, it will take years).

By the time we're sure we can implement a JVM without getting our asses sued, people might well have moved on to different languages (I know I'm considering it).

Re:Oracle is doing everything they can to fuck up

[JAlexoi]

Yet, if you fork OpenJDK(witch is GPL) you get all of the patents in there per GPL. GPL v2 and patents
Passing the TCK is required for you to use Java trademark.

Re:Oracle is doing everything they can to fuck up

[modmans2ndcoming]

missing 2% that is irrelevant to 98% of development is not something to really care about when talking about it being an alternative.

FOSS-Nuts would extol the virtues of Kaffe before Java was opened up and no one said a peep about how utterly horrible it was and how much it was missing. Mono is nearly identical to .Net and for the UI, GTK# easily offers a cross platform toolkit.

Going GTK# means you get designer support in monodevelop. It is a very good tool.

Mono is a soon to be feature complete c# framework that is nearly identical to .net. a little redevelopment of UI code is worth it to many people if they are trying to sell to move to a cross platform market, including mobile devices.

Performance-tuned Java?

[kaaona]

Wouldn't that be like racing whales?

Re:Performance-tuned Java?

[Stevecrox]

Java isn't that much slower than C++ these days, if you do it right Java/C++ performance is so close as to not matter.

It's also more maintainable, has better frameworks and you don't have lots of beginner/intermediate level programmers introducing memory holes left, right and center.

Saying all that I work for a company which has invested millions into Java applications. Considering how Oracle has been acting the tech leads are pushing to moving us back to C++.

Re:Performance-tuned Java?

[Cornwallis]

Very good. I was going to ask about the idea of premium JVM vs regular JVM. How could it be tuned to perform any worse than it does now?

Re:Performance-tuned Java?

[zlogic]

Hear, hear. I've developed an image processing algorithm in Java and C++ (pretty simple: for every pixel in floating point array, compute some basic stuff, create a few classes to simplify the storage of temporary values and save the result into another array). The code was as close as possible in both languages, with no obvious screwups like memory leaks or unnecessary copying of stuff. To my surprise Java ended up being 15-20% faster than C++. And C++ is THE language for image processing, every new image processing algorithm is written in C++ (with the occasional exception of C) because of performance reasons.

Re:Performance-tuned Java?

[hvdh]

Now try using SSE intrinsics. With Java, you can't do that. In C(++) you should get a nice speedup ending up several times faster than Java, unless you're bound by memory bandwidth.

Re:Performance-tuned Java?

[bertok]

Now try using SSE intrinsics. With Java, you can't do that. In C(++) you should get a nice speedup ending up several times faster than Java, unless you're bound by memory bandwidth.

Or use a better C++ compiler, like the Intel one, which gives a substantial speed boost with no developer effort.

There's also ready-made C++ maths libraries for pretty much everything, many with SSE optimizations.

These micro benchmarks also ignore cache thrashing, which kills Java performance for large apps.

Re:Performance-tuned Java?

[Tridus]

I keep hearing this from people, but I've never seen a fast Java app. Is it just that anything done in Java with a UI runs like a beached whale, or is it just that most people who make Java UIs are doing it wrong?

Re:Performance-tuned Java?

[Anonymous Coward]

Or you can use Mono.Simd.

Seriously, Mono these days seems like a much saner and more open framework than Java.

Re:Performance-tuned Java?

[devent]

Or maybe you just talking crap? Eclipse, Netbeans, Freemind, Visual Paradigm, Lotus Symphony. All have a lot of users and I'm using Eclipse all the time. For GUI apps I couldn't care less what language is used, for them are 99,99% of the time waiting for user input. In fact, sometimes Firefox is way slower then Eclipse.

What people don't realize is that only because an application is written in C or C++ you don't need the same amount of code and abstractions that you have with the JVM. You need memory management, exceptions, logging, threading, etc. You need it all in a "native" application, and the JVM is a "native" application anyway (it's written in C++). So what you have in Java is a tiny bit more abstraction, which gives you so much more productivity for the developer.

Re:Performance-tuned Java?

[Nursie]

"if you do it right Java/C++ performance is so close as to not matter."

Then the vast majority of java apps I've encountered were done wrong!

Re:Performance-tuned Java?

[Lennie]

That is the whole problem: if you do it right

I see to much shitty programming though.

Re:Performance-tuned Java?

[swilver]

Startup time? I don't think I would notice the difference when it is well below half a second.

[~]$ time java -jar /data/Backup.jar
error parsing commandline: Missing required option: [-c continue the last backup, -i create a new incremental]
usage: Backup [options] databasefile target dir|files...

  -c,--continue continue the last backup
  -h,--help display this help
  -i,--incremental create a new incremental

real 0m0.163s
user 0m0.131s
sys 0m0.022s

Re:Performance-tuned Java?

[Lennie]

(And it took years to get it to perform in anyway sane).

Re:Performance-tuned Java?

[Lennie]

I meant to write: any way

Re:Performance-tuned Java?

[TheRaven64]

Or use a compiler that supports autovectorisation. Actually, this is where it gets more interesting. In theory, the Java version can be faster, because Java lacks pointer arithmetic so the compiler / JVM is free to rearrange the memory layout to better suit the vector unit. In practice, I'm not aware of any implementations that actually do this.

Re:Performance-tuned Java?

[TheRaven64]

Google uses Java a lot on their servers

Concrete example please. I've only visited Google a few times, but they made it pretty clear that C++ and Python were the two languages that they used for pretty much everything.

Re:Performance-tuned Java?

[Spikeles]

you don't have lots of beginner/intermediate level programmers introducing memory holes left, right and center.

Hehe, sure sure, just wait till they start adding action listeners to things, you'll be leaking memory in no time.

Re:Performance-tuned Java?

[arendjr]

Are you dynamically allocating memory during your calculations? Basically new and delete are pretty slow in C++. While garbage collection is slow as well, actually allocating of memory is much faster in Java. Fortunately, you can implement your own allocation strategy in C++ by overriding the new and delete operators. Admittedly, it's a bit more work but can in many cases easily result in a tenfold speedup.

Re:Performance-tuned Java?

[Kjella]

I know some enterprise backends that run in Java with a web frontend and that seems mostly fine, the run like a 747 in that they're huge, resource intensive but take a large number of passengers and are reliable. I think most of java's bad reputation comes from the UI classes. They look like crap, they are memory hogs and slow. And it seems like no java developer has ever learned the pattern with one GUI thread and one worker thread to do the work so it doesn't freeze up and shit.

Re:Performance-tuned Java?

[hvdh]

I tried GCC (iirc 4.2) and ICC 10 autovectorization on an existing code base of about 50 filters. Each filter is available in plain C code and using SSE intrinsics.I fed the plain C versions to both compilers. GCC could only autovect the simplest filters. ICC vectorized several more, but not even half of them.
Even where autovect worked, the results were slower than the hand-written SSE intrinsics code.

Re:Performance-tuned Java?

[Lennie]

Was that with or without it being in the filesystem-cache ?

Re:Performance-tuned Java?

[GreatBunzinni]

If you want your anecdotal comparison to have any meaning then you must at least mention what environment you used to run your Java code and what compiler you used to test your C++ code. Moreover, wirting code which is "as close as possible in both languages" means that you've written code in at least one of those languages which sacrificed efficiency for the sake of being "as close as possible" to the other one.

So, in essence, what it means is that your anecdotal evidence is meaningless.

I am tired of getting the occasional java fanboy boasting Java's miraculous performance when compared to any other language, including C and C++. Yet, once they descend into the realm of reality they suddenly are faced with the fact that Java is is slower and hogs up memory when compared with the C++ equivalent compiled with GNU g++, no matter what benchmark you run, no matter what VM you run your benchmarks on. That means that Java fanboy's do the language a disservice by boasting Java's fictitious performance excellence. Java is a great language and it does have it's place but once you try to shoehorn it into domains where it clearly is not the best tool for the job then you are quickly faced with problems and limitations.

Re:Performance-tuned Java?

[modmans2ndcoming]

much saner and pretty much complete through .net v4

Re:Performance-tuned Java?

[Tridus]

And as a user of a program, I give a shit about the productivity of the developer... why?

I care about the productivity of the user: me. Eclipse is a lot of things. Fast is not one of them.

Hell, the worst example is probably SQL Developer, which is an Oracle product. In the same time it takes that thing to launch I can go get breakfast (or start Visual Studio, load a project, and be doing work). Bizzarely if you have the VS tools for Oracle installed (also an Oracle product), you can fire up VS, connect to Oracle, and load/edit procedures far faster then you can do it in Oracle's Java based tools.

Now maybe that's just Oracle sucking. But Flash Builder? Not exactly a speed demon either (and that's built off Eclipse). Hell, even something small like Gallery Remote is slow.

So yeah, maybe these miraculously fast Java UI's exist somewhere... or maybe your definition of "fast" isn't the same as mine. I'll take something written in C++ any day as a user.

Re:Performance-tuned Java?

[PastaLover]

I don't suppose this counts for you, but it's worth point out that google app engine code is either Python or Java.

I do know at least one blogger (Steve Yegge) that used to use java at google and has done some work on tool support. At least all his blog posts on how much java sucks seem to suggest it. :-) The google testing blog (before it became shit) has in the past also focused a lot on java-specific testing techniques, suggesting they do a lot of java.

Re:Performance-tuned Java?

[PastaLover]

As a java coder, when someone says "Java app" I almost never think of anything with much of a GUI (unless it's a web page). The swing libraries are notoriously bad and a real weak spot in the entire framework. The problem is mostly that they're not just badly designed (not with performance in mind, and usually lousy bindings to the underlying platform), but also that they're hard to work with. Considering most coders actually coding up GUIs really don't know what they're doing, it would help if it was all a bit more straightforward.

I dunno what AWT is like, but from personal experience with Eclipse I'm not all that positive.

Re:Performance-tuned Java?

[gtall]

When you say Java isn't that much slower, I presume you mean non-gui apps. I've not looked at Java recently, but is Swing on the various platforms still running like a slug compared to native apps?

Re:Performance-tuned Java?

[jeremyp]

And as a user of a program, I give a shit about the productivity of the developer... why?

Because the developer will produce software with more features and fewer bugs sooner and more cheaply.

Re:Performance-tuned Java?

[HyperQuantum]

Have you tried to find out why that is?

Maybe it's because in Java you can allocate new objects quickly, without having to 'destruct' them right after they go out of scope or something. And the compiler might have been able to optimize more because of the constraints of the language (less possible aliasing etc.)

Re:Performance-tuned Java?

[owlstead]

I think there is a lot of things said for using Java over C++, most of which count because of faster development time, more *compatible* libraries but also indifference to CPU used. What I never understood from Java is that it is *too* conservative regarding which features to support. If I would setup a framework, I would make it more modularized, plugable, so that developers *can* use all the hardware features of a particular platform. After that you make sure that "Java compatible" applications always have a generic way of doing things - e.g. the way that OpenGL also runs in software.

Just depending on the features that are fully generic over all systems is taking it a bit far, and I think that is the main reason why Google could not opt for Java ME (at least from a technical point of view).

Re:Performance-tuned Java?

[devent]

And there is again the half-knowledge of Java. First, you are free to make such "plugins" with JNI, which you can write C/C++ code and use it in Java. That is how SWT works, the OpenGL Java implementation, and the extensions on MacOS for SWT/Swing. You are free to write your performance critic code in C/C++ and use all the SEE you like.

Second, JavaME have already a ton of extensions, and have profiles. That is the problem with JavaME, that there is no one JavaME for all mobile phones but more or less forked JavaMEs for each mobile phone. Google is successful with Android, because you have only the Android platform and can develop vendor neutral.

Last, the common ground with JavaSE or JavaEE is very much needed and appreciated so as the developer you can say that your application will behave the same on different platforms.

Re:Performance-tuned Java?

[devent]

You mean one of those Visual Studios that took over 5 minutes to load a project? I have used VS2008 and VS2010 and both took way more time to startup and load a project on the same machine as my Eclipse. Furthermore, I didn't had any plugins for VS installed, so it didn't tool longer but it was less useful as Eclipse, too.

The overall GUI was slower then Eclipse, too. So I'm really don't know why the people post on forums that Eclipse is slow or whatever and I don't have a high-end computer either. Maybe you can make a video so I can see it for myself?

Re:Performance-tuned Java?

[oiron]

If it's that close, either your algorithm isn't that heavy, or you're doing it wrong. My (anecdotal, of course) experience is that there's a significant difference if you do things right. Any void pointers or heavy casts in the C++ version? Try templating it instead...

Also, with c++, you can do things like SSE which are not possible with Java, and depending on your algorithm, can give you a hell of a boost.

Re:Performance-tuned Java?

[oiron]

Strange. If I write a trivial hello world in Java and in C/C++, total execution time for the C/C++ one is less a hundredth of the Java one, with its lumbering great JVM dependency to initialize. Even shell script or Perl beats Java dramatically.

Be fair - Hello World is hardly a good test of performance... Shell script is obviously highly optimized for string processing, and Java is more generic. The question is, how much of a speedup do you get between JVM initialization and good optimization in C++? What's the tradeoff?

For example, if the JVM overhead is 10 seconds across the board, and your Hello World takes a second to run, that's a massive slowdown. But if your program is a complex image processor that takes 10 hours to run, the JVM overhead shouldn't be noticeable. Again, the JVM may have a lot of optimized methods within the standard library for certain things (say search) which are faster than what you hand-code in C.

Of course, you can probably write a faster method in C, but you have to know what you're doing. If you don't, or if the JVM method is already optimal, you won't gain much, and may actually end up losing a lot. But again, see my above post - in general, a really optimal method written in C/C++ should normally either equal or beat the equivalent JVM method hands down.

Re:Performance-tuned Java?

[Anarke_Incarnate]

Now make this a trading app where every 20 microseconds are an eternity. Do you care then?

Re:Performance-tuned Java?

[turkeyfish]

Excellent point. If Oracle can't get its Java PR under control and looses the larger Java community, who begin to bale out of Java, then one will likely see lots of JNI like hooks or plugins to accomplish the intrinsics associated with performance issues as becoming an increasing alternative to the JVM, even if these are not strictly compatible within a single build model (ie different executables for different platforms).

Re:Performance-tuned Java?

[fishbowl]

>And as a user of a program, I give a shit about the productivity of the developer... why?

In business software, the user of the program is often the one waiting for the developer to finish, and is paying for the time.

This thread took a direction all about user interfaces and algorithm performance. Computationally intensive algorithms tend to be pretty rare in business software, and the user interface often amounts to a report, or a web-based UI. Imperative logic, simple internationalization, abstraction of relational data, and a framework that lets you forget about the concerns of deployment or the web or service interface are all more important that runtime performance or 2D or 3D graphics.

I realize that there is stuff out there to compete with enterprise java (I happen to like Django, and I also think PHP has a place in this world), but for a lot of business development, it's pretty hard to beat e.g., Spring and Hibernate, and using that stuff means using Java, and implies a whole bunch of things about Java in the middle tier.

Re:Performance-tuned Java?

[careysub]

Wouldn't that be like racing whales?

Whale racing! That sounds so cool! I'd definitely pay to see that!

Would the various whale classes be defined by size or species? Do larger whales have an advantage (in the long distance categories I would think so)? I can imagine a whale-racer importing a larger southern hemisphere blue whale as a ringer to compete with their northern hemisphere cousins.

Re:Performance-tuned Java?

[Stevecrox]

True, but in my limited experience people who write inefficient Java code, would write inefficient C++ code and people who write bad code will write bad code in any language, there are more bad Java programmers than C++ because its easier to learn (pointers).

The most common problem I've seen with inefficient Java code is iterating through arrays and lists. That problem wouldn't go away just by using C++.

Java has things like Checkstyle, PMD, Findbugs, etc.. which are free and can be integrated into your IDE. These tools provide management metrics and force proper practices making it easier to do it right. I'd love to know of some C++ equivalents.

Re:Performance-tuned Java?

[EsbenMoseHansen]

Ok, but to be fair at least Eclipse and netbeans feels like a beached whale. Then, most IDEs do.

Admittedly, the Java straightjacket has never been my cup of tea in any case. Nothing personally, I think every programming language I have tried just sucks for no good reason. For Java it would be: no operators except the build in ones, no static asserts/polymorphism, no proper resource management, no closures and general uglyness of a lot of constructs like "switch" and "if".

Re:Performance-tuned Java?

[Wrath0fb0b]

Fortunately, you can implement your own allocation strategy in C++ by overriding the new and delete operators. Admittedly, it's a bit more work but can in many cases easily result in a tenfold speedup.

Please don't do this. I work in a performance intensive field (physical/biological simulation) and every smart*** that's tried this has failed miserably or succeeded for a very modest (TCMalloc (google) or HOARD (uMass). Both are quite good, especially when dealing with heavily threaded applications.

If those don't solve your problem, then maybe you can think of starting to tinker with your own memory allocation. Maybe.

Re:Performance-tuned Java?

[ADRA]

Maybe, but I probably wouldn't be restarting the application every fraction of a second either.

Re:Performance-tuned Java?

[Celarent Darii]

Any project with more than a few files is often painfully slow in Eclipse. Especially if you are editing the gui classes. Startup is better now but it can be painfully slow. Eclipse might be convenient, but it has never been fast.

Even Java apps like JEdit take an enormous time to launch with comparison to a "native" application such as TextMate.

Java is more than a bit of an abstraction, it is truly a virtual machine, a sort of emulator that will always be slower.

Re:Performance-tuned Java?

[FoolishOwl]

Saying all that I work for a company which has invested millions into Java applications. Considering how Oracle has been acting the tech leads are pushing to moving us back to C++.

The super-rich gamble enormous amounts of wealth at their fancy gambling tables, wealth which takes years of hard work and ingenuity to create. They profit from the fallacy of the broken window.

This is sad. These are exactly the sorts of dilemmas from which FLOSS is intended to save us.

Re:Performance-tuned Java?

[Circuit Breaker]

Are you aware of a single Java compiler that supports vector operations at all? Either autovectorization or intrinsics without JNI would do, but I'm not aware of any.

Re:Performance-tuned Java?

[martin-boundary]

Please don't do this. I work in a performance intensive field (physical/ biological simulation) and every smart*** that's tried this has failed miserably or succeeded for a very modest (TCMalloc (google) or HOARD (uMass). Both are quite good, especially when dealing with heavily threaded applications.

On the contrary, controlling memory allocation in an app is very important/useful. I don't think the parent was talking about replacing the underlying malloc() calls with a homebrew malloc(). He was likely talking about allocating a large chunk of memory once(ish), and accessing/managing the contents as a custom memory pool specially tuned for the app. That's what placement new() is really for.

Re:Performance-tuned Java?

[TheRaven64]

To be honest, the only times I pay attention to JVMs at all is reading papers from IBM (mainly their T J Watson group) for some ideas to steal^Wadapt for ObjC and Smalltalk implementations. I'm not aware of any, but that doesn't mean that they don't exist. The VM has the opportunity to do run time profiling of loops, so if I were writing a JVM it's one of the things I'd do.

In terms of intrinsics, you probably wouldn't get anything worthwhile because they'd have to be sufficiently general to use on all Java-supporting platforms. You'd want to expose something a bit higher-level to the JVM, like a vector class with all of the standard java.lang.math things supported and implemented with either vector ops or fall-back scalar code. If anything, autovectorisation would be easier...

Re:Performance-tuned Java?

[petermgreen]

if you do it right Java/C++ performance is so close as to not matter.
The thing is java's language design seems to put design for performance and memory efficiency at odds with design for readability.

Specifically Java encourages you to create lots of tiny objects each with a seperate heap allocation, a pointer to a virtual method table, a pointer pointing at the object and probably other overheads too.

You can avoid this by various tricks but those tricks will all make the code less readable.

C and C++ on the other hand encourage you to create arrays of structures which have virtually zero overhead since they are allocated in a single block of memory and afaict C++ doesn't create anything like a virtual method table unless you specifically use things that need it as part of your object.

Another thing to consider when benchmarking is that GC changes WHEN the price for your actions is paid. With traditional memory allocation/deallocation you pay for cleaning stuff up as part of the process of using it. With GC you pay for it as a seperate GC process that happens at times outside of your control (either when the vm thinks it has time to spare or when it hits it's memory limit, whichever comes first).

Re:Performance-tuned Java?

[Javagator]

"as close as possible in both languages" means that you've written code in at least one of those languages which sacrificed efficiency

I agree. For example, the most efficient way to allocate small transient objects in C++, is to put them on the stack. Something you can’t do in Java. I love Java for portable, non-cpu intensive applications. But if you really need to run fast, C++ is the preferred language.

Re:Performance-tuned Java?

[Anarke_Incarnate]

You have not met our programmers :)

Re:Performance-tuned Java?

[JonySuede]

please do a paper

Re:Performance-tuned Java?

[exomondo]

The code was as close as possible in both languages

What's the point of that then? Write an optimal version for each language, use language-specific features, use the best-suited compilers, optimisations, etc... Different languages require different approaches and different tools, achieve your result by actually using the language features, libraries, etc... not just whacking the same code in both.

If you're seeing that much of a performance difference in something so basic there's a good chance you're doing it wrong in one of the languages, obviously there's a good chance sub-optimal c++ will lose out to optimal java.

Re:Performance-tuned Java?

[EsbenMoseHansen]

Eclipse uses SWT, another toolkit not part of the standard lib. It sucks. I have never used AWT, but that was the predecessor to Swing...

The language itself is partly at fault, since it GUI programming means handling a lot of resources (window handles e.g.), but Java is notoriously bad at handling resources bar memory.

Re:Performance-tuned Java?

[PastaLover]

Eclipse uses SWT, another toolkit not part of the standard lib. It sucks. I have never used AWT, but that was the predecessor to Swing...

D'oh. Yeah I meant to say SWT. I have written a GUI or two in AWT. It makes Swing look good, I think that says enough.

The language itself is partly at fault, since it GUI programming means handling a lot of resources (window handles e.g.), but Java is notoriously bad at handling resources bar memory.

I personally think the class design for Swing was overly ambitious and probably just too constrained by their desire to reuse at least some of the AWT stuff. SWT provided a clean break, but apparently that didn't help matters much. I'm not at all sure that the problem is really inherent in the language as such, but these things are hard to quantify.

It's also not as if a decent UI layer actually needs to be wonderful and fun to work in. I write some JSP now and then and it's painful, but it works (though btw, also can have huge rendering times considering what you're doing is outputting some text to a page). I've heard people gripe about how awful coding their UIs in ObjC (+cocoa?) can be, but they get it done anyway. I feel the error was in trying to provide a "pure" class framework to do the work, instead of some hack that actually works.

Re:Performance-tuned Java?

[EsbenMoseHansen]

The problem with Java re GUIs is that Java does not have a reasonable way to handle resources. Try writing a snippet that copies a file to another, and remembers to close them both :) Then recall that in a GUI, you'll have dozens of resources to manage all the time. That makes it harder.

IF I was going to do GUI in Java, I'd look at QtJambi, even though it is no longer officially supported. But that is because I know Qt/C++, which is rather pleasant to work with, even if it sometimes takes a bit of exploring to find out how something is done.

Re:Performance-tuned Java?

[PastaLover]

The problem with Java re GUIs is that Java does not have a reasonable way to handle resources. Try writing a snippet that copies a file to another, and remembers to close them both :) Then recall that in a GUI, you'll have dozens of resources to manage all the time. That makes it harder.

I don't know, aside from all the exception handlers it's pretty much just C, right?

Re:Performance-tuned Java?

[EsbenMoseHansen]

Pretty much. With C, you usually handle resources by using goto (cleanup_step_n).. which Java does not have. With C++, of course, you have several good options, which all are more or less RAII in some form.

Re:Performance-tuned Java?

[PastaLover]

Java has try-finally, with the obvious downside that you need to wrap everything in there and you go at least two levels deep for something like a file copy (i.e. try-except wraps two try-finally's). JDK7 or 8 should have a with keyword or something of the like, which will hopefully help. You definitely have a point that it isn't very readable. I'm more concerned here though with how they access OS/window system primitives (i.e. windows etc). That might be a cause of a lot of our pain.

OTOH, other VM languages have the same problem, and they seem to do just fine.

Re:Performance-tuned Java?

[EsbenMoseHansen]

Most other VM machines have closures, which can also work. Java is pretty unique in having no RAII, no goto and no closures. At least, this was my experienced in the 3 attempts I have made at getting Java and GUI into the same sentence.

As an aside: all languages that have objects really should have closures... it is really just a matter of syntax at that point. At least C++ will *finally* get that now.

Can you Digg it?

[Yuioup]

Java has become the Digg of languages.

Goodbye Java, we hardly knew thee...

Sun monetized Java too

[trifish]

They bundled the spyware Google Toolbar with it (optional, but opted-in by default in the installation options).

Re:Sun monetized Java too

[nitehawk214]

They bundled the spyware Google Toolbar with it (optional, but opted-in by default in the installation options).

Actually its the Yahoo toolbar, which is equally as annoying.

Re:Sun monetized Java too

[KarmaMB84]

I think they were offering a random toolbar. I'm pretty sure I'd seen everything from MSN, Yahoo, Google and Bing over the years installing Java....

Uh, watever, just migrate to Python, Perl6, Lua...

[polemon]

It's out of question, that this will kill Java as preferred language in academia and science.

But who cares, really? There are other languages, that would be a more than adequate "replacement" - if I may call it that - for Java.
So professors will have to teach Python in university, how is that something bad?

Java was chosen a few years back, because it was modern and cross platform, but that is Python as well. I also suggested using Lua in academia. For teaching programming and data structures, this is arguably one of the favorable languages.

I'm a Perl developer, so I'll wait and see what happens with Perl6...

Don't jump to conclusions

[jernejk]

Oracle already has free and pay-for JVM: HotSpot is free, JRockit is not. I expect the free JVM will be just fine for desktops and small servers. I'd expect pay-for JVM to target enterprise solutions. And again, I expect them to ship this JVM for free with their middleware products (Weblogic etc.). But yes, this sucks for JBoss.

Re:Don't jump to conclusions

[jonwil]

No reason why JBoss team couldn't put some effort into improving OpenJDK to make it faster or better performing or whatever is needed to be up there with other J2EE solutions.

Re:Don't jump to conclusions

[modmans2ndcoming]

considering the state of JBoss... I don't think they would succeed.

Re:Don't jump to conclusions

[CynicTheHedgehog]

Why? I've been using JBoss for years and it seems to be a very robust, agile, and well-managed project, particularly after being acquired by RedHat.

Besides, I wouldn't expect "JBoss" itself to mess with the JVM. I would expect RedHat and SuSE to put effort behind maintaining/optimizing the free JVM, especially given their respective stances on free software.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Ruke]

That's a nice list of scripting languages you've got there. And don't get me wrong, scripting languages are nice. However, if speed is an issue, Lua's never going to cut it in the same way that Java does.

We'll give you the first bytecodes for free...

[Anonymous Coward]

... But after that it's going to cost you.

"With Oracle it doesn't have to make sense. It just has to make money." -James Gosling.

ME and EE

[leuk_he]

Don't forget that there are other edition going round as well for a long time,

The Java MobileEdition (that was never free, and caused the fields of use clause in their license) and the Enterprice Edition (license?). They did use different classes, but i think the same vm(correct me if i am false). Makeing a differnce in vm is not a big step.

Re:ME and EE

[jernejk]

I'm not sure about ME, but EE uses the same JVM, just different flavor (-server switch in JVM). EE is implemented by application servers, not JVM.

Re:ME and EE

[Vectormatic]

ME does not use the same VM, as a matter of fact, ME is still stuck at the feature level of java SE 1.4

Re:ME and EE

[hlavac]

EE is just a set of APIs on top of SE. It's not a different version of Java.

Wow....

[mark-t]

How can the people at Oracle not see that this sort of maneuver will only _decrease_ Java's popularity?

Why did Oracle buy Sun in the first place, exactly?

Re:Wow....

[headLITE]

I don't get it. JRockit was always proprietary. Why should they make it free just because they have the good sense to consolidate their JVM projects into one?

Re:Wow....

[IllusionalForce]

Why did Oracle buy Sun in the first place, exactly?

They're clearly trying to be the biggest trolls on the entire Internet. Ever.

Re:Wow....

[ChunderDownunder]

Agreed, provided the core JDK remains contributed to the openjdk project under the GPL, this discussion seems tainted by FUD.

JRockit historically sped up BEA's Weblogic and it sounds like nothing more than Oracle's existing offerings benefiting from a pluggable interpreter that uses JRockit. The difference being it'll re-use more of the existing VM codebase.

Hotspot itself exists in several flavours inc 'client', 'server' and the community contributed 'Shark' based on LLVM. To this Oracle will soon add JRockit in the premium binary-only release.

Re:Wow....

[TheRaven64]

Why should they care? Java's popularity is not something that makes them money. If it doubles the number of people paying Oracle to use Java and decreases the number of people using Java but not not paying Oracle by 90%, it's a net win for them.

Re:Wow....

[Teckla]

How can the people at Oracle not see that this sort of maneuver will only _decrease_ Java's popularity?

At this point, too many companies are too deeply invested in Java to switch. It'll be cheaper for them to write Oracle big checks than it will be for them to port their millions of lines of code to other languages and platforms.

What Oracle is doing well and truly sucks for smaller businesses and individuals, though.

Re:Wow....

[gman003]

Agreed, provided the core JDK remains contributed to the openjdk project under the GPL, this discussion seems tainted by FUD.

No, this discussion is tainted by blind panic (and maybe the ability to see trends). We know what happened to OpenOffice, we know what happened to MySQL, so we're assuming the same sort of things will happen here. It's really only FUD if someone is actively spreading fear, uncertainty and doubt about the project.

Re:Wow....

[jeremyp]

Unfortunately, Java's popularity does not make them any money. In fact, it's a burden because they have to provide industrial grade servers to cope with the downloads and the bug tracking and they have to pay all the people who maintain that stuff as well as the dev team of course.

Oracle got successful by selling extremely expensive database servers to not many people. That's the business model they understand. And it clearly works better than giving away the product for free to everybody and hoping some of them would buy a Sun server to run it on. That's why we are talking about what Oracle will be doing with Java and not what Sun are going to do about their newly acquired database server.

Re:Wow....

[Stan Vassilev]

I don't get it. JRockit was always proprietary. Why should they make it free just because they have the good sense to consolidate their JVM projects into one?

Tech-inclined people like to complain how their non-technical bosses are easily swayed into bad tech choices.

It seems tech people as just as easily swayed into silly conclusions by fact-free ideology and hearsay.

To me, this article defines a very good future for Java, to both those using the open JDK, and those needing more in the enterprise. I'm sure once the echo chamber gets tired of their "Oracle is evil" chant, they'll figure out they've spent all this time screaming wolf over nothing.

Re:Wow....

[gtall]

Yep, and it ironic that Oracle benefited from Sun giving Java away freely. Oracle (and MS) live by the Breshnev Doctrine: what's mine is mine and what's yours is open to discussion.

Re:Wow....

[Anne Thwacks]

Java's popularity does not make them any money.

Well, D'oh. Java is popular Because its free.

Once its not free, its not going to be very popular. Never mind speed, or any other feature. As for the C#, Mono people: some people just dont like to put their neck in a noose. call it superstition if you like, but I call it being over 21.

Its not a bad thing.

[drolli]

I think it will be more about supported configurations etc. in that way it will be easier to swallow depending on a jvm in the core of your product. Having things like "the java update broke sth on a certain configuration" not answered by "congratulations, you problem" IS a plus.

Sid you mean Java or Java-VM or Java-SE or Java-EE

[krischik]

I agree with this statement, this is indeed the birth of new opportunity - A new technology to replace Java...

If you want to replace Java you can just use Scala. It can do all that Java can do and is a lot better. Or did you mean to replace the Java-VM, or Java-SE or Java-EE. Then please be more precise in future.

Opensource and corporate control bad combination

[Eelco]

Over the years I've seen lots of companies trying to combine the open source development model with a for profit (and publicly funded) business plan. While this seemed okayish at the time, this whole Oracle debacle is clear proof that the opensource development model combined with some corporate entity controlling it is risky from an opensource perspective. Big opensource projects hopefully have learned from this and go the Debian route : turn yourself in a NGO or something and never worry about shareprice or corporate takeovers again.

It is a real shame that big projects like MySQL and OpenOffice are in this position. Maybe it was more or less opensource, it sure as hell was not independant. Sure forking works, but recreating the entire organisation (and funding) will keep you from developing quite some time.

Platform jump

[krischik]

A language jump is easy: use Scala. What you mean is a platform (Java-SE, Java-EE) jump.

Re:Platform jump

[ubersoldat2k7]

Tell me, how is Scala supposed to work if, for some reason, some of the features it uses, go to the pay-to-use JVM?

Re:Platform jump

[ls -la]

I believe that was the point he was trying to make.

Re:Platform jump

[krischik]

As "ls -la" said: That is the point I am making: That is that most posters say Java and really mean JVM or Java-SE.

Re:Platform jump

[ADRA]

1. I hate functional languages, so scala looks and feels generally against my grain. I'm not the only one, and trying to shoe-horn people toward the wrong solution will not win you any points
2. Even Gosling found Scala to be complicated and you expect the regular rank and file corporate developers to adopt it?

Scala, Groovy, Ada.

[krischik]

We already got better languages like Scala for the JVM. The Ada for JVM project has been restarted as well. If it just the language you want to replace - that is not a problem.

So what do you want to replace here? Ahh, well as you are an AC you won't answer.

Re:Scala, Groovy, Ada.

[wmac]

The language you are pushing in multiple messages will never gain enough ground.

Re:Scala, Groovy, Ada.

[Anonymous Coward]

We already got better languages like Scala for the JVM.

Yay, let's push a language that's even more bent on allocating trillions of disposable objects per second than Java.
But no, that's not a problem, because if you reserve 2GB of RAM per application to amortize the speed cost, it's not that slow!

Re:Scala, Groovy, Ada.

[swilver]

...and how would it effect those languages when there's only a sub-standard free JVM available?

Re:Scala, Groovy, Ada.

[TheRaven64]

Yes, because allocating short-lived objects in a modern JVM is a very expensive operation. No, wait, it's an increment on a pointer value stored in a register. Disposing of them is marginally more expensive, but if only very slightly. The cost is roughly equivalent to allocating a C++ object on the stack.

Re:Scala, Groovy, Ada.

[QuoteMstr]

You made your point. In context, "Java" clearly and unambiguously refers to the entire ecosystem. Give it up.

Re:Scala, Groovy, Ada.

[Yvanhoe]

It is not about having a better language, it is about having a language with a larger sales department. Face it : most language use decisions in companies are made by clueless bosses once the salesman from a few proprietary companies have met him. So it will be either Oracle's new and fancy java ("he said it will be really easy to develop, faster than ASM and work cross-platform without any work") or Microsoft C# or a variant for the .Net platform ("they said it is better than Java and they just drag and drop to do GUI, it HAS to be a powerful language")

Re:Scala, Groovy, Ada.

[Anonymous Coward]

That's exactly what I was talking about. You people are content with the fact that Hotspot has pretty fast memory management, ignoring the fact that it needs a significant area of memory reserved just for being able to bump that pointer without triggering a collection every couple milliseconds.
Oh yes I know, RAM is cheap. Hardware is cheap for corporate applications, desktop and mobile apps need to make due with what the user has.

Re:Scala, Groovy, Ada.

[140Mandak262Jamuna]

Frankly, the fact that you think the GUI builder is part of the language or the platform is laughable.

OK This is slashdot. You dont read the original article. But don't you read the comment you are replying to? The GP is not saying "GUI builder it part of language". GP says decisions are made by morons who could think so.

Re:Scala, Groovy, Ada.

[Haeleth]

You are confusing the theoretical cost of ideal garbage collection with the actual cost in a particular implementation.

I have worked on optimizing real-world Java applications that really were running too slowly. The problem really was that they were allocating too many short-lived objects in a modern JVM, and reducing the number of allocations really did improve performance significantly. Sorry if reality doesn't match your fashionable assumptions, but that's just the way it is.

Just look at some benchmarks some time. Scala performance is closer to Python than Java. Yes, often that's fast enough. No, it is not always possible to throw hardware at the problem when it isn't.

Re:Scala, Groovy, Ada.

[jbengt]

Face it : most language use decisions in companies are made by clueless bosses once the salesman from a few proprietary companies have met him. . . . ("they said it is better than Java and they just drag and drop to do GUI, it HAS to be a powerful language")

As far as I can see, there's only one way to interpret it, and that's that the poster was quoting the legendary clueless boss, not stating his own opinion.

Re:Scala, Groovy, Ada.

[krischik]

Actually Scala was just the example I like most. I try to push the idea that there is a difference between Java, JVM and Java-SE and most people get it mixed up. Note the OP of the article got it right: He speaks of an improved JVM and mot an improved Java compiler.

Re:Scala, Groovy, Ada.

[krischik]

Not quite. It is not a language decision. It is a platform / framework / ecosystem decision. But I guess the pointy haired bosses you are talking about don't know that as well.

Re:Scala, Groovy, Ada.

[krischik]

If you say so. But then original article talked about an improved JVM and JDK and not an improved Java. So I still think it was an important point to make. No matter which definition of Java you use.

Re:Scala, Groovy, Ada.

[FranTaylor]

Don't forget that you actually have to call the constructors, too.

Re:Scala, Groovy, Ada.

[master_p]

You forget the cost of going through the collectable objects and invoking their finalizer. The cost of the gc is not only allocation and deallocation, it's also the cost of scanning the graph, relocating objects, updating pointers, restoring ghost references, and invoking the finalizers.

Re:Scala, Groovy, Ada.

[shutdown -p now]

You forget the cost of going through the collectable objects and invoking their finalizer.

Vast majority of Java objects do not have finalizers (i.e. don't override Object#finalize), and VM is perfectly capable to figure that out and don't call it for them. This is even more true for short-lived objects - those have finalizers pretty much never.

Re:Scala, Groovy, Ada.

[aug24]

Out of interest, did you do the analysis separating out the actual gc cost, or just lumping it with creation/destruction?

malloc/init/free isn't cost-free in any language, so any language will get a benefit from re-use if it can be designed in, whether it is a gc or manual memory management language. I'd be surprised if the majority of the cost was gc, but am happy to be proven wrong. Every day's a school day ;)

Re:Scala, Groovy, Ada.

[dgriff]

I think you'll find an awful lot of short-lived objects these days are allocated on the stack. Modern JITs use escape analysis to figure out whether the object ever "escapes" from the current stack frame. I.e. is the object still referenced when the current method returns. If it isn't (which is often the case with short-lived objects), the object is allocated right there in the stack frame (just like alloca) and there is zero GC cost. Part of the beauty of Java is that you can do this kind of escape analysis because everything is so well-defined.

Re:Scala, Groovy, Ada.

[Nethemas the Great]

People can throw up all manner of alternative languages as equal or superior but the problem isn't the language per-se. It's the fact that there is a massive ecosystem built around and on top of Java that doesn't exist for these alternatives. Oracle knew this, and other things and will be leveraging them to their full potential.

With Oracle no matter what their venture, it's always been about the money. They believe that their acquisition of Sun will net them more than their initial investment. They could care less about the long term health of Java so long as they can apply strategies to bring them ROI. Thanks to some beautiful foresight on Sun's part with the GPL and patent protections, Java for the time being is safe. I do not however, believe this will continue indefinitely. Oracle's task now will be to make what is available, and possible with open-source Java as limited and marginalized as possible so that moving forward the only reasonable choice will be to purchase their non-free offerings.

At this point it's difficult to predict where things will go. I think that the sector most in doubt at this point will not be the desktop but rather the enterprise. Java is a well established platform in the corporate world, and given the present investments of IBM, Red Hat, the Apache Foundation, etc. in enterprise Java I believe we'll see it continue on. However, with the innovation fracturing from the J2EE specification instead of being sown back in as it has in the past. Instead of being partners in innovation as was the case with Sun they will now be competing against Oracle. Whatever happens though, I think it is pretty much a given that the Java world will not flourish as it did when Sun was at the reigns.

Re:Scala, Groovy, Ada.

[JAlexoi]

As always, high performance it's not a general statement it's always an application and environment specific topic.

OS support for C#

[krischik]

But C# needs CLI and .NET as platform. And those platforms don't have the same OS support as the JAVA-VM has.

Of course most here mix up language (Java) and platform (Java-SE, Java-EE) and don't really know what they want to replace with what.

Re:Wow...

[Anonymous Coward]

NOKIA will be assimilated by Microsoft. The process is already underway.

Meanwhile, at Microsoft...

[rennerik]

... Ballmer et al are wringing their hands nefariously as they see the future of C#'s marketshare increase by leaps and bounds. And that's good for Microsoft in every way, since every application written in C# instead of Java means a license for Windows is being purchased to run each copy of the software. In web apps, it's a server license; in workstation applications, it's a desktop OS license. Either way, it's a win-win for Microsoft, and a massive loss for Oracle.

Not that I mind, per se. I prefer C# in every way to Java... but from Oracle's perspective, I don't see how they see this would do anything but hurt Java and their reputation that's rather ubiquitous.

Now if only Mono would get their asses in gear and not lag so far behind .Net versions, there would actually be an open source OS alternative to running modern C# applications.

Re:Meanwhile, at Microsoft...

[tsj5j]

Actually mono isn't lagging THAT far behind. .NET Framework 4.0 was released in April, mono added compatibility for it in September.
Of course there are still some missing classes, but considering Microsoft had a 2.5 year head start, mono is actually doing pretty well.

Re:Meanwhile, at Microsoft...

[modmans2ndcoming]

Mono is pretty much caught up. will all features of .net

There are a few exceptions, but nothing that can't be worked around quickly.

Re:Meanwhile, at Microsoft...

[PJ6]

But what's sad is, with Java you can actually host a real application in your browser. I'm still waiting for .NET to have something similar (Silverlight is a joke).

Re:Meanwhile, at Microsoft...

[No. 24601]

... Ballmer et al are wringing their hands nefariously

Ya, I'm sure they are. But too bad I see this as more a win for the true open source development and dare I say even Stallman et al. and his philosophy. Anyone with huge investments in Java should have had second thoughts when it became clear that Sun wasn't going to "donate" its control of the language, its libraries and its VM-licensing to a non-profit industry-run consortium which should have been setup as early as ten or even fifteen years ago.

This is not a win for MS at all for a simple reason: once bitten, twice shy.

Re:Meanwhile, at Microsoft...

[shutdown -p now]

Of course there are still some missing classes, but considering Microsoft had a 2.5 year head start

Mono is still completely missing WPF, which was released at 1.0 four years ago in 2006 - and for which several betas were public available even earlier. They don't have any specific plans to implement it, either.

Re:Meanwhile, at Microsoft...

[BitZtream]

Considering Mono is still missing 2.0 methods, which it STARTS implementing a framework version is really irrelevant isn't it?

Framework support isn't exactly Mono's only issue either.

Re:Meanwhile, at Microsoft...

[Ash-Fox]

Mono is pretty much caught up. will all features of .net

There are a few exceptions, but nothing that can't be worked around quickly.

None of my J# applications work in Mono and after repeated queries for help on the matter on forums, IRC channels, I gave up. I don't see this "work around quickly" method you speak of.

Re:Wow...

[Nursie]

WTF??

Seriously, WTF are you on about? Have I missed something?

That would be terrible!

Consistent

[mseeger]

At least they are consistent: first they killed OpenSolaris, then they managed to split the OpenOffice community and now they will marginalize Java. I am sure they have something in store for MySQL too...

CU, Martin

Re:Consistent

[VendettaMF]

Don't forget the killing off of multiple smaller projects and developments.

Admittedly, till now the only one I've cared about was the Sun Game Server (Project Darkstar), but I'm sure there'll be plenty more.

After all, it's not like Oracle can even produce a decent database of their own any more.

Re:Consistent

[coredog64]

They've already laid out their cards on MySQL: The cost of support for MySQL just doubled this week.

Re:Consistent

[mseeger]

I know.... But my guess is that they need more measures to kill it.

Re:Consistent

[BitZtream]

They don't have to do anything with MySQL. The fear of them pulling the rug out from under it at any point in time is enough to slow or reverse its adoption by larger companies.

Some insight from one of the bigger customers...

[Anonymous Coward]

Working in a senior role within a global investment bank, we buy a lot of vendor product, especially from what is now Oracle (Oracle Databases products, Weblogic products, etc.) - and if they want to charge us for the 'better' JVM going forward, no doubt we will pay for it. As will the other banks.

And Oracle knows this. It does not give a shit about small-scale Java customers, but the big corporates, like us, well, they know that even if we decided tomorrow that all new projects were to move to C#, or C++, or Objective-C, or whatever, that it would take a long time to change course, and Oracle can still bill for a long time.

One thing to remember - our bank gets and stays profitable because it pushes a lot of IT outside to third-parties (offshore developers are *much* cheaper than in London and NewYork), and they do not see any problems with getting a global price agreement with companies like Oracle and Microsoft.

Personally, I am brushing up my C++, learning Objective-C and C#, as I think the medium and smaller companies in the market will start to migrate away from Java, as the cost savings of cheaper Java developers is lost once you have pay large amounts for the Java install and licensing.

Stallman wrote the Java trap, and we all laughed. Sun is nice we thought, it'll be ok. We were all wrong. Stallman saw further, he saw that even if Sun was ok, if someone bought Sun, then things could get messy. Welcome to messy.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Anonymous Coward]

However, if speed is an issue, Lua's never going to cut it in the same way that Java does.

Correction:

http://shootout.alioth.debian.org/u64/benchmark.php?test=all&lang=all&d=data&calc=calculate&gpp=on&java=on&luajit=on&v8=on&lua=on&tracemonkey=on&box=1

Re:Surprised?

[Lennie]

It has for example been revealed that Oracle explicitly asked about Google and patents before the deal and supposedly didn't even want the deal if it wasn't useful.

Re:Wow...

[contra_mundi]

The new Nokia CEO is Stephen Elop, the former head of Microsoft Corp.’s business unit.
Source.

Also there have been talk of and/or more integration of their services:
https://www.microsoft.com/Presspass/press/2007/aug07/08-22NokiaMSLiveServicesPR.mspx
https://www.microsoft.com/presspass/press/2009/aug09/08-12pixipr.mspx
http://techcrunch.com/2010/09/15/will-microsoft-and-nokia-team-up-to-take-on-apple-google/

Re:Wow...

[Nursie]

Ah yes, thanks, I did hear about the new CEO.

Guess we'll have to watch this one carefully. Nokia was in need of a shakeup, but with an MS man at the top I fear for the future of Meego on Nokia phones.

Re:In a way that's reassuring.

[Lennie]

Vala ?

Legacy

[sgt101]

1000's of big companies (telcos, utilities, retailers, gov, defence) use java in their back office, and... well everywhere.

This may cause them to change their policy for new software development, and it may also squeeze the java developer market badly, but for sure there will be strong arguements for splashing £50k here, £90k there, £20k somewhere else, on getting the new JVM to pick up the performance of application x, y, z which are long in the tooth and a pain in the arse.

The alternative is to rebuild, which carries risk - although would be a good move in the long run. In the meetings someone will say "yeah, but we are all dead in the long run" and that's that basically. As a CIO you just pay over £50k, get your users back on side, keep your job for another year, collect your bonus, put another years pension contrib into the pot.

So, Oracle will make money, lots of money, off this. You guys can squeak, MS will cheer, the Python community will see a boost (perhaps), but Larry and co will be richer.

Mysql (in the future) = Oracle feather light (down load it and run it and you are up and going in less than 1hr - oracle normal = 6hrs to setup?) But, if you are an enterprise DBA then you want the management and recovery features that Oracle gives you (as well as the scaling - even though it gets so mind bendingly expensive).

Open office - who cares?

Oracle bought Sun to be IBM mark 2. Expect them to buy Accenture next.

Re:Legacy

[Lennie]

Are you trying to say Java will be the next Cobol (lots of legacy code and a small expensive workforce) ? Or is that just a bridge to far.

Re:Legacy

[modmans2ndcoming]

Java was the next Cobol about 5 years ago.

Re:Legacy

[gman003]

Considering how many classes are taught in Java, the "small expensive workforce" bit is out. "Lots of legacy code", though, might be right. Java may end up being the language that everybody knows, but hates.

Re:Legacy

[sgt101]

TBH. COBOL is very easy to learn, it's jsut a pain in the arse to build anything substantial in it because it's very primitive in comparison to a modern language.

COBOL experience is rare because the details of the kit that underpins it (DB2, forms for terminal input) and the interfaces to that kit are difficult to access; you can't spend 3 months doing an open source project on COBOL and then talk about that in your interview.

Java is everywhere - and will continue to be accessible as a technology, just much less desirable in corporate architecture if it is vendor owned in the way that this announcement predicts. So I can see a move away from Java for new projects and a need to support the old stuff falling to a less skilled and poorly paid workforce.

Re:Legacy

[shutdown -p now]

Now it is the current COBOL.

Re:Some insight from one of the bigger customers..

[Lennie]

It is funny how you mention global investment bank and Stallman is right in the same post. :-)

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Anonymous Coward]

Yes, because Lua sure doesn't cut it for performance-critical applications like computer games.

http://en.wikipedia.org/wiki/Category:Lua-scripted_video_games

What's the big deal

[Anonymous Coward]

Back in 'the day' JRocket (BEA) was more feature-rich and faster than the Sun JVMs; and had a pricetag. I don't remember anyone complaining about BEA. "Grr. How dare a company introduce a JVM for profit! Corporate bastards!" How is this any different? Why can't Oracle have a superduper JVM they charge money for and have a free one? There's an infinity of software companies that operate with this model. Also why are you angry with Oracle? You should be angry with Sun for mis-managing it's assets and living in business-model-fairytale-land which resulted in their having to sell out. Thanks to Oracle Java didn't 'go down with the ship' and all the Sun employees didn't lose their jobs.

Re:What's the big deal

[gerrytucker]

I was just about to post the same point. From a large company perspective, the purchase of the premium JVM will most likely get you different tiers of support. For example, if you have a critical software project and for some reason the JVM is not playing nice with your project you could ring up support. The free JVM would allow you to submit bug reports I'm sure, that will then get rack and stacked in whatever priority Oracle chooses. Paying for support gets you much better, guaranteed support.

Also, this seems just like the model that larger Linux distros have been taking. Anyone can get the free version, but to get premium support and additional "features" from the companies you pay for it.

Re:What's the big deal

[Courageous]

How is this any different?

It's different, because in this case it's the same company controlling both the product that everyone uses, and the pay edition. There is suspicion that they will stop adding new features to the free edition in order to incentivize folks to buy the pay edition. Essentially, it's a trust problem.

C//

Re:What's the big deal

[yuhong]

I suspect many Slashdotters commenting did not do their research on the history of the products.

Alternative JVMs

[janwedekind]

Well, Google did just that and it's called Dalvik. Who cares? Oracle does!

Re:Uh, watever, just migrate to Python, Perl6, Lua

[vbraga]

LuaJIT has pretty amazing results.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[modmans2ndcoming]

I would be so happy if academia stopped using Java. I hate Java but need to use it for every freaking development project for my Masters. arg.

Google and Android

[Compaqt]

O would that Google had licensed Java Mobile Edition from Sun, given Sun that additional amount of revenue, and thereby allowed them to stave off takeover attempts instead of using tricks to use Java/not-Java without paying for it on Android.

Re:Google and Android

[metamatic]

Have you tried developing for Java ME? If Google had taken that route, Android would suck, and Apple would own the smartphone market.

Re:Google and Android

[SanityInAnarchy]

Given that Android was meant to be free and open source, requiring people to buy Java ME first would kind of undermine that. It's like accusing Mozilla of trying to save money by not implementing H.264.

Re:Google and Android

[Compaqt]

Just a very little bit (how about you?). I've kind of been getting my toes wet with it with Netbeans Mobility Edition and Sun Wireless Toolkit.

There's a graphical editor to add controls. You set validation for fields. You can set the navigation graphically. Seems OK to me, but I'd love to hear your opinion on it (and other platforms) as well.

Also, it's hard to see what it is you can't do with J2ME.

Graphical games? If you can make Bubble Bash2 or Rally3D, you can probably make any game you want. Netbeans even has a game wizard with sprite editor.

You can access phone contacts, the file system, bluetooth, and so on with J2ME. You can use the camera, tuner, music/equalizer, GPS, the (inter)network, 3d graphics, video, crypto, SMS, and the touch screen.

Re:Google and Android

[Compaqt]

Given that the Android device makers are paying out $10-15 (?) per open-source handset to Microsoft, could they not have paid a few dollars to Sun? I'd say they probably ought to have paid something (anything) morally even if they didn't legally have to.

Or was Sun just the codependent code-donor that everyone loved to abuse while fearing and paying out to MS?

Now if you say there's no way a corporation can justify giving free money like that because of fiduciary duties, etc., then I'd have to say they deserve everything coming to them from Oracle. Because by that very same moral standard, they can't justify leaving money on the table.

Re:Google and Android

[SanityInAnarchy]

The point is that right now, no one needs to do that. Right now, any handset with sufficiently open hardware, I can install my own version of Android without paying anyone.

So, what are they paying Microsoft for? Office compatibility?

Re:Google and Android

[alannon]

J2ME is a joke. It's more or less limited to the core libraries and language constructs of a 12-year-old implementation of Java with the addition of an almost laughable lowest-common-denominator set of UI widgets. The language and core API limitations mean that it's impossible to use any but the most trivial of the incredibly rich ecosystem of open-source Java libraries out there. There ARE API extensions out there that add back support of some of the modern Java libraries, but they're provided piece-meal by different mobile venders, offering no real useful target for cross-platform development.

Android, on the other hand, supports almost all of the language constructs of the latest Java 1.6 as well as an extremely rich portion of the modern Java core APIs. From this, an enormous amount of existing Java code can be leveraged. On top of that, it provides a rich platform-specific set of APIs. Yes, it will only work on Android, but at least I can use most of the tools and libraries from the modern Java ecosystem.

Cross-platform UI design on even very similar platforms (graphically, at least) such as MacOS and Windows is still extremely difficult. Mobile devices are far more divergent in their UIs than desktops. J2ME doesn't really solve anybody's problems by pretending to be a complete cross-platform solution.

Re:Google and Android

[hotfireball]

I guess, metamatic won't answer you anymore. :-)

Re:Google and Android

[mcvos]

Yes, but Google didn't want Java ME. They wanted Java SE on a mobile device, and Sun wouldn't allow that.

Nothing new, it's not like JRockit was free

[Anonymous Coward]

How is this news? It's not like JRockit VM was exactly available for free:

http://www.oracle.com/technetwork/middleware/jrockit/downloads/faq-085599.html#t2

I see this as being more of the same - if everyone's panties didn't get in a bunch every time some entity tries to make a $ off of something then maybe you'd understand that.

So, while I'm no fan of Oracle how about we all just shut up and be happy that the JRockit features are going to be merged with already excellent HotSpot VM.

If Oracle wants to add "special sauce" to the JVM so that thier stack can offer more to thier customers then so be it so long as it doesn't impede the open source JVM.

Re:Nothing new, it's not like JRockit was free

[yuhong]

Nope, because the open source Hotspot makes it redundant!

It's platform, stupid !

[boorack]

For me Java/Oracle problem is more about platform than it is about language. In the old times there were platforms, like UNIX, Windows or Netware. OS/Hardware combo was a deployment platform. Now we have JVM, .NET and web browser - these are the new major platforms (plus niche things like Erlang BEAM, Parrot, Ruby/cPython interpreters, to some extend LLVM).

Let's take JVM as an example: you have defined instruction set (bytecode), well defined ABI (this one is much better than in conventional operating systems) and well defined set of standard services (standard libraries). You also have class loader which somewhat resembles dynamic linker functionality in the conventional OS. Oh, and there is a pretty damn good debugging/profiling/monitoring infrastructure built in. And from application programmer point of view JVM is pretty much like a OS. Programmer can use many languages to target this platform, not just Java. It is possible to implement almost any language on top of JVM (albeit some things have no practical sense, for example C/C++ with its pointer arithmetics).

Would Larry prove its intent to totally screw Java (I'm still not sure of it yet), we'd need to have another platform rather than another language. There are enough cool languages to choose from, but aside from JVM and CLR there are no viable, widely supported multi-language, multi-paradigm platforms. JVM is propably the best one available but as it ages, there are more and more shortcomings visible. Having enough support from companies and developers (and from Larry screwing up Java) one can design and implement a new VM addressing some additional things, like:

- native support for dynamic dispatch (albeit OpenJDK7 seems to support it in some way) - what I mean by that is trying to achieve performance somewhat comparable to statically typed programs (now we mostly compare to C implemented couterparts, eg. JRuby vs. Ruby, Jython vs. cPython etc.);

- support for big memory heaps - most VMs suck at this (except for Azul), so we have to slice server machines and run many instances of JVM on one machine, then cluster/farm these JVM which is both silly and troublesome;

- better support for massive concurrency - again, most JVMs suck at this and Java thread model isn't perfect and isn't suitable for everything;

- support for multiple independent garbage collector zones - some language may utilize this to mitigate concurrency/big memory heap problems (Erlang, anyone?); ability to use different garbage collection algorithms in different zones if it makes sense (ex. big heap as in Java vs. small heaps as in Erlang);

- ability to execute on multiple target devices at once - to utilize GPUs/APUs directly from bytecode (maybe with some limitations), without those crappy hacks we see today; it also applies to memory management that seems to be a horrible hack in current GPGPU solutions;

- better support for long running VM processes, mainly hot code loading (Sun JVM sucks at this but some other solutions like JRockit seem to do a better job), maybe some code versioning, better tools to administer / tinker with running VM process (similiar to what Erlang has);

There is more than 15 years since Java was published. There is about 10 years since Microsoft built its CLR. And there is a lot of new things that appeared (GPUs, huge memories, multicores) and lots of new knowledge we obtained since then (look at all these JS interpreters in modern browsers!). There is also a pretty good base to build on (LLVM, V8, BEAM, PyPy and tons of other projects). On top of such VM we can implement various languages (including Java), maybe even better than JVM.

With enough help from friendly enough companies (RedHat? Google?) we can propably do much better than JVM and leave Larry and his corporate cronies in the dust. As long as there is a good quality reference implementation we don't need to chase Java APIs nor we do need to beg for TCK access.

Re:It's platform, stupid !

[CarbonShell]

Amen, brother!
I'd love for a Java replacement (my suggestion: BINJ, BINJ is not Java)

Only thing we would have to watch out for is the 'feature creep' many languages die to.
f.i. scala: normally you'd think 'academic' is something bad, but people complain that it is 'not academic enough'.
or 'not expressive enough' ... meaning 'let's make it MORE cryptic because I am to flipping lazy to actually write code'.

Re:It's platform, stupid !

[Tim C]

Would Larry prove its intent to totally screw Java

I'm not sure quite what you mean by "totally screw", but if you mean "kill off" that's not going to happen any time soon. Oracle has a huge amount of middleware that is either written in or depends heavily on Java and the JVM. At worst, they'll make it painful or expensive for other people to use - but then that would be exceedingly foolish, as driving developers away from your products isn't good business...

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Anonymous Coward]

Interpreted languages, such as Perl, Python, Lua, Ruby and whatever pet scripting language is trendy at this moment, do not nor they should have any place in any academic curriculum. Let me explain.

  First of all, a higher education courses which are based on programming languages should be designed to educate the students on fundamental topics such as programming paradigms and design patterns. It is unthinkable for a higher education course to waste both the teacher's and student's time doing vocational training. After all, if you are enrolled in a higher education program then you should be smart enough to handle that stuff by your own. Llet the teachers teach the important technical and scientific concepts which are harder to grasp.

Then there is the reason of applicability. If a course is based on a technology which doesn't have a stable basis and which is likely to drastically change due to a project leader's whims then it is certain that the code written in that particular technology will suffer from bit rot and that your skills which you earned throughout your course will essentially become meaningless. In short, due to a technology's dynamic nature, those who invest their time and effort in a degree or course which is based on that single tool will soon find out that they have wasted their time.

Then there is the issue of control. If a language is not standardized then you don't have any assurance that your code will run at all in the future. This issue is particularly problematic if a language does not provide any specification regarding the implementation or is based on a single reference implementation whose project leadership often invests their time and energy in complete rewrites.

So, knowing this, languages such as Python, Perl, Lua, Ruby or whatever trendy language you just picked up is not nor it ever could be sanely adopted by a higher education degree. Doing so is a recipe for disaster and it does a disservice to all students.

Business as usual

[elygre]

But, but... that's the way it already is!?

Today, Oracle offers a free JVM, which is the one they got from Sun. They also offer the JRockit JVM which has "performance tuning and tie-ins with Oracle's middleware". It's a "premium" JVM, and you have to pay to get it. If anything changes, it's the fact that they want both of them to work off the same codebase, bringing at least some of JRockits premiumness into the commodity JVM.

Move along, nothing to see, except a world full of distrust and hostility.

Re:Business as usual

[owlstead]

Yep, basically this is an old story being played over and over again. The big question if these kind of sentiments will screw up the Java community enough to migrate away.Of course, I do presume that most people won't read either Slashdot or the Register. Certainly, if they do, they should be warned ad nauseum about the way both tend to report on certain kinds of matters.

In the case of The Register you could substitute that with "on all kind of matters". Basically the way The Register reports is that it takes an early stance on an issue, and then they make a mockery of their enemies in all the follow up articles. This makes for fun reading, but they should have chosen a black and white color theme instead of red and white for their front page.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[gbjbaanb]

Except...

once Java didn't cut it in the same way lower-level compiled languages did. But then memory became cheaper and processors faster, and Java started to become fast enough. (I recall my PC having 32Mb Ram when Java first came out)

Today memory is even cheaper (I have 4Gb) and processors are even faster (4 cores!) which means that those scripting languages can cut it where VM-based languages used to be needed. Slashdot runs fine, and its in perl after all. At work, I'm considering performance issues on a 24 core server with 48Gb RAM. (.net issues this time - not really down to .net, but the design of the code more than anything. It could have been written in script and the result would be exactly the same)

So, the problem isn't that scripting languages are crap, too slow, they are - but the hardware has caught up to the point where everything you need to do for the usual LoB app can be written in them without anyone noticing that they are. We have Mercurial SCM written in python, and rich GUIs written in javascript. Unless you have some HPC requirements, script is fast enough.

Or to put it another way, Java is now obsolete.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[PastaLover]

The engine for those is usually written in C++.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[gtall]

Does Scala allow me to build user interfaces? If not, it really isn't a replacement. What I need is something that allows me to build cross platform guis.

Monetize?

[nurb432]

Didn't you really mean marginalize? Not that this wasn't expected.

Re:Some insight from one of the bigger customers..

[obarel]

In other words, Java will become the new COBOL - large corporations will still invest in it, because it's cheaper than replacing everything. But no startup would touch it because they can see it has no future.

I don't really mind - I still program in C ;-)

yawn

[Zecheus]

There has always been a free and premium jdk from oracle.

We don't have enough information in this announcement about the difference between free and premium in technology, performance or price.

This announcement is much less important to java enterprises than it is to java detractors.

Microsoft was better option?

[cloudcreator]

I wonder what would MS do with Sun...

Re:Uh, watever, just migrate to Python, Perl6, Lua

[RocketRabbit]

Ironically, this is similar to what people were saying about Java when it was new. Oh, it's cute and good for little whizzy web apps, but it will never cut it on the server.

Well guess what Matilda, as machines get more powerful people program in higher level languages.

Re:java fork

[owlstead]

You get bashed by patents, that is what happens. Personally I don't see *any* benefits of it being open source the way they play it out now.

No Way In Hell...

[littlewink]

will you see any significant move from JVM to .NET. Java users will move to open-source languages instead.

Why move from one proprietary language to another? This discussion has already been heard.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[mfnickster]

Can't you use Qt to build your GUIs? It's available on Windows, Linux, Mac OS and mobile platforms.

Scala runs on top of Java anyway.

How so?

[Sycraft-fu]

Last I checked, Sun wasn't going out of business. They didn't go bankrupt and have Oracle buy up all their stuff at a bankruptcy auction. They were doing fine, and Oracle said "Hmm, we'd like to get in to those markets, we are going to buy you."

Remember that any public company that isn't majority owned by someone attached to it is subject to a buyout. If the majority of stock is out there in public hands, another company can simply buy up enough of it to buy the company. Usually it is far more cordial than that, they contact the big shareholders and work out a deal, but it can be done in a hostile fashion too.

So that Sun got bought doesn't mean they weren't doing fine, it just means that someone bigger than them decided they wanted to own their market.

Re:How so?

[turkeyfish]

You obviously aren't familiar with what transpired.

Sun stock went into the toilet with the .dot com crash and McNeely spent more time talking a good game than in developing a viable business strategy by failing to diversifying away from SPARC or making SPARC good enough to make it worth the premium price. Their Java efforts turned out to be misguided as a means of accomplishing the latter, since it only emphasized that from a customer perspective there was little premium to be had by buying SPARC. Schwartz came on board too late to steer a different course, particularly as th tech economy was like the rest of the market in a tailspin. Board members like McNeely, who were near retirement age anyway, decided to sell out knowing it was the only way they would get that golden parachute they had been dreaming of. Towards the end as is usually the case, you saw more and more of Sun's profits directed toward big exec bonuses as they prepared to sell out, insuring the ultimate death of the company as a viable independent business.

Microsoft investors should be getting nervous about Ballmer's recent announcement of sale of 1.2 Billion in stock. This is how the stock market works these days. Its an inside game played by insiders, while boilerplate fantasy is sold to the public and the poorly informed.

Re:How so?

[Raenex]

This is how the stock market works these days. Its an inside game played by insiders, while boilerplate fantasy is sold to the public and the poorly informed.

No need for the "these days" qualifier. The stock market has always been an insiders game.

Re:How so?

[OdinOdin_]

Re: Ballmer's recent announcement of sale of 1.2 Billion in stock

Maybe he just sees the US markets and the Dow Jones is still due for a large correction (in the coming years) and that he does not think he will himself still be in the driving seat by the time global economies pick up again. I do not see this is 'trouble at Microsoft' just a wealthy man protecting his wealth through diversification.

After all the US is now leading the race to the bottom with the printing press (Re: QE2), if I were Balmer I too would want to be moving my wealth out of the USA, since everything is about be become more expensive there (this is the side effect of cranking that printing press, the world won't let you escape what is owed, raw materials/assets will be become more expensive to soak up that money, which in turn diminishes your ability to export your way out of the hole).

Re:How so?

[drerwk]

From Wiki:

In 2007–2008, Sun posted revenue of $13.8 billion and had $2 billion in cash. First-quarter 2008 losses were $1.68 billion; revenue fell 7% to $2.99 billion. Sun’s stock lost 80% of its value November 2007 to November 2008, reducing the company’s market value to $3 billion. With falling sales to large corporate clients, Sun announced plans to lay off 5,000 to 6,000 workers, or 15-18% of its work force. It expected to save $700 million to $800 million a year as a result of the moves, while also taking up to $600 million in charges.

Loosing 80% of your value is not doing fine. And no, mergers are not all about owning the market, if that was a real risk the merger would not have been allowed due to anti-trust reason. If you disagree, then tell me what market Oracle bought. The purchase happens for a public company when the purchaser offers the stockholders more than they think the current management can return to them. If Sun had been doing fine, they would have been too expensive to buy.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[zhong-guo-1]

Well. I could whip up some cross platform guis in Scala, but the real question is: Can it web scale? If it doesn't have web scale synergies, then lets be real - it's no better than perl.

Re:f1rst

[gagol]

You are right, this is about Oracle screwing itself... but that joke is getting old.

Recording Industry Math...

[Junta]

I've heard it time and time again when some piece of software owned by a large company is open source or at least given away for free as a tool to sell other stuff or addons.

Inevitably, some business leader will see that the company 'moves', let's say, 10 million installations of product at 0 dollars. They will inevitably pull some number out of their ass as to how much they think that software is worth (say 100 dollars per install, though invariably they'll come up with a far more convoluted per-client,per-instance, per-core scheme). Then, they'll say 'if we had only charged 100 dollars like it is truly with, we would have made a billion dollars!'. They'll fight, debate, capture the imagination of someone actually empowered to make the call and poof, start trying to charge 100 bucks a pop for a formerly free product.

Suddenly, the rate of new installs drops to a trickle at best, leaving everyone wondering 'what happened?' when the answer should be obvious to anyone with a cursory understanding of basic economics.

Of course, it may be that Oracle fully realizes this is a long shot, but they somehow think that without this panning out, they see no point in Java really surviving anyway. As a technical observer, that seems like a horrible mistake on the level of cutting their legs off with respect to their commercial Java software, but who knows.

Re:Some insight from one of the bigger customers..

[gtall]

Stallman was talking out of his ass but playing the probabilities. He didn't see a Sun trap, he calculated the probability of a Sun trap was significant. Unless you want to think of Stallman as Nostradamus...come to think of it, there is a slight hairy resemblance.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[gtall]

Yeah, I could if I wanted to code in C++ (rather eat a broom) and do my own memory management...although I think Qt has some support in this area.

I guess I'll have to bite the bullet and jump back to C++...yeeeaccchhhhh!

Re:Wow...

[Anne Thwacks]

with an MS man at the top I fear for the future of Meego on Nokia phones

Dont worry, with an MS guy at the top, Nokia has no future anyway. They had already lost the plot, with a whole bunch of phones, none targeted. The lack of a N900 successor, and third party Meego devices means Meego is already a dead duck. Buy an HTC Desire, install LeeDroid on it, and love it Sure you have to root it, but you are a nerd, and rooting is just a step to keep the plebs down.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[oiron]

But with Unladen Swallow, Python just might...

And for any really serious performance considerations, jump to a real language that you can actually optimize in (C, C++, etc).

Re:Sid you mean Java or Java-VM or Java-SE or Java

[gtall]

Synergies? You aren't a real programmer, are you? Just kidding, sounded like you swallowed a marketdroid sometime in your past.

I wasn't interested in web scaling a gui, but I'll admit that is an important consideration for some. Isn't that segment of the market getting taken over by Javascript, html5, and all the other weboria crap? I just wanted a good cross-platform gui language that gives native looking app and maybe requires a few tweaks here and there to make it work like native app...like what I had to do for Java.

When I used Java, I used the AWT. Swing was too new and seemed to suffer speed problems. Sun never got small machines. So, their graphics objects got allocated for every damn object on the screen in the paint call chain when all that was needed for most was merely the correct bounding box. Once I circumvented that, it sped up quite nicely.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[SanityInAnarchy]

JRuby can do that (just by calling Swing directly), so I imagine Scala can, to.

The problem here is the JVM, not Java the language.

Nah, Oracle will make a nice bundle out of this.

[williamyf]

Nah, if the slow (Glacial anyone?) migration of IE6 corporate apps is any indication, this will work great £€$¥ wise for Oracle.

A free Java VM, no mater if it is performance or feature crippled means that anyone with a quick and dirty need, and the java skills (prevalent nowadays) can solve it with a quick java fix, instead of, say, learn another language, or risk it with a comunity derived JVM (You know, and I know and mabe our hypotetical programer know that the comunity derived JVM is technically top notch, but most likely his PHB and the C?O people do not).

When the application gets ingrained in the internal business process, and you need something (performance or feature wise) found only in the paid for VM, the corporate sweet daddy will be strong pressed to pay for it, rather than write the app from scratch in a more open language.

The only way this would backfire is if one big player like IBM, HP, or SAP adopts one of the comunity derived forks and starts offering paid support (and the associated credibility) for it...

Sad but true.

Of course, this only applies to corporate internal use sw. The shrink wraped SW ((F)OSS or not), and the movile/mobile sw have different dynamics...

Re:Uh, watever, just migrate to Python, Perl6, Lua

[onefriedrice]

That's a nice list of scripting languages you've got there. And don't get me wrong, scripting languages are nice. However, if speed is an issue, Lua's never going to cut it in the same way that Java does.

Actually, there is no intrinsic property of scripting languages that make them slower than Java. Obviously it mostly depends on the implementation of the virtual machine:

http://shootout.alioth.debian.org/u64/which-programming-languages-are-fastest.php

For example, take a look at this list. LuaJIT actually seems to be quite competitive with the fastest of the languages. So, no, I can't agree with you. If speed is an issue, your choices may be limited to those languages which have just-in-time compilation for your processor type, but it certainly doesn't mean that Lua et al. can never match the likes of Java.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[mario_grgic]

Yes, Scala has full access to Java APIs including Swing.

Re:Wow...

[SanityInAnarchy]

Why would I want to "keep the plebs down"? I want them to become more educated, more self-sufficient, and more nerdy, so I have to do less family tech support.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[mfnickster]

I haven't actually done any Qt programming myself, but the demos I've seen were nice.

It looks like it has bindings for other languages than C++ :

http://en.wikipedia.org/wiki/Qt_(framework)#Bindings

Re:Uh, watever, just migrate to Python, Perl6, Lua

[LaurensVH]

That's a nice list of scripting languages you've got there. And don't get me wrong, scripting languages are nice. However, if speed is an issue, Lua's never going to cut it in the same way that Java does.

Have you seen LuaJIT really? It's not exactly slow. (Also, what the hell is a 'scripting' language?)

Re:Sid you mean Java or Java-VM or Java-SE or Java

[negRo_slim]

It looks like it has bindings for other languages that C++:

If he needs that much hand holding I think it's a lost cause bro.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[melonman]

> Interpreted languages, such as Perl, Python, Lua, Ruby

I'm not sure about the others, but Perl isn't an interpreted language. The script is compiled, and then the compiled code is executed. That's not a million miles from a JIT compiler.

> is trendy at this moment

Maybe I missed somethng, but wasn't Perl running half the Internet before Java was a twinkle in Gosling's eye? There are many things to be said about Perl, but "new-fangled" isn't one of them.

> It is unthinkable for a higher education course to waste both the teacher's and student's time doing vocational training

Right, but why is that an argument for Java? You can teach Java as a way to tick CV boxes, and you can teach other languages as a way to really understand programming.

> If a course is based on a technology which doesn't have a stable basis and which is likely to drastically change due to a project leader's whims.

You're saying that the Perl community rushed into Perl6? Or maybe that a decade-long upgrade cycle for C++ standards is too rapid? And that the future JVM roadmap as of today is very clear to you?

> Then there is the issue of control. If a language is not standardized then you don't have any assurance that your code will run at all in the future.

Not sure why this is an issue for education but, in any case, backwards compatibility in Perl is legendarily good. There are plenty of production sites running Perl code that is decades old.

I can see an argument for teaching CS using C or assembler (because they force you to understand what a computer does under the hood). I can see an argument for teaching CS using functional languages such as Haskell. I can see an argument for using Perl because it's possible to demonstrate a much wider range of programming styles than with Java. ("Whatever the question, the solution is an object" doesn't seem to me to be ideal in an educational context.)

But I can't see offhand why Java is different in kind to any of the other languages in your list as far as suitability for education goes. It's not a very pure implementation of OO programming, it doesn't teach you anything about memory management. If you are not careful, you give students the impression that it's impossible to write code without an IDE. Java has its merits, its fans and its detractors. But it's not self-evidently The One Right Language to use to teach CS.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

Scala runs on top of the JVM and support both Java-SE and Java-EE and even Android development. I don't suggest the later but for the former two it is pretty good.

So yes: You can write apps for the AWT, Swing and SWT GUI with Scala.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

Scala is fully Java-EE compatible including Servelets. Does that answer your question?

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

Indeed.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[gtall]

No, I don't need any hand holding, I've probably started building apps (guis, realtime embedded, network drivers, educational logic proggies, etc) since before you were born. I was just lamenting that C++ programming isn't an uplifting experience, especially when it comes to guis. But then you probably see nothing wrong with writing network drivers with it.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[gtall]

Yes, but then I still must have a JVM around. The point was that if Snoracle is going to fuck up Java, then I cannot rely on it and must look elsewhere.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[MillionthMonkey]

It's a fucking compiler. The object code looks like it could have come out of javac [ideally]. You can work with the same classes and run compiled scala/java/groovy/closure/etc. code through the same EE bullshit.

What's a "web scale synergy" anyway? How many do you need? I would think Perl had at least one puny web scale synergy to show for itself.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[theshowmecanuck]

Come on. The article is about the monetization of the Java VM. If your critical reading sucks that much I'll make it easy, he means replace the Java VM. That means Scala is not a solution. Please use your fucking brain in future.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[Profane MuthaFucka]

Parent poster has seen Qt, but has not used Qt. Possibly hasn't used Boost either. But he's going to flap his ignorant gums anyway.

My fellow geeks, ignore him.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Haeleth]

Ironically, this is similar to what people were saying about Java when it was new. Oh, it's cute and good for little whizzy web apps, but it will never cut it on the server.

Yes, and then JVM technology improved to the extent that Java performance is within an order of magnitude of C++ for most common server tasks.

Java was faster than Perl/Python/Ruby to begin with, and now it is vastly faster. Meanwhile every new popular scripting language actually seems to be slower than the last. And time and time again a startup tries to use something like Ruby or PHP on their servers, runs into massive problems, and ends up either sinking vast sums of money into trying to compile PHP to native code, like Facebook does, or completely rewrites their website in a JVM-based language, like Twitter did.

Well guess what Matilda, as machines get more powerful people program in higher level languages.

No, it is more that the higher-level languages are getting more efficient. C++ became popular when C++ compilers got good at optimizing C++. Java became popular when the JVM got good at JITting Java. Hobbyists and academics have always used high-level interpreted languages -- Python today occupies much the same niche as BASIC did in the 1980s. It's got less to do with hardware than you think.

The hardware argument made sense briefly in the 1990s, when machines suddenly got powerful enough for it to be briefly possible to consider writing everything in an interpreted language. But now it's the 2010s, and we are increasingly moving to less powerful mobile devices with limited battery life, and looking for ways to shrink our servers to reduce the financial and environmental cost of supplying all those machines with power and cooling.

Welcome to the modern world, where suddenly efficiency matters again. If your favourite language is stuck in the 1990s, maybe you'd better start learning something else.

Cut it out

[FranTaylor]

with the ad-hominem "this is going to kill java" stuff.

Oracle has said openly that there will always be a free JDK available from them. We also have IcedTea which is not going to go away.

Oracle is going to sell and support an enterprise version to their enterprise customers, with enterprise features. Stuff that you probably don't use on your desktop or even your small server. Stuff like infiniband support and enterprise message queueing. If you need this kind of stuff, you are either already an Oracle customer or a potential Oracle customer. Otherwise you will be perfectly happy with the free JDK and nothing is going to change in that department.

So cut it out with the cringing and moaning, it's pointless.

Google Supporting Open Source Java?

[careysub]

You do realize that Google uses Java extensively ?

Indeed.

And it raises the question about whether Google is going to safeguard its own software investments and create a Java development team for open source Java large enough to provide the core development expertise and resources formerly provided by Sun.

Self interest suggests that they should be taking the lead role in Java development simple t support their own business.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[RocketRabbit]

Wait, are you seriously asserting that language tuning is the sole reason that computers are faster these days? It has nothing to do with the vast speed increases in the CPUs and buses? If you take a modern Java and stick it on an original 1995 era Pentium, I think somehow it will be even slower than Java 1.0 was, as the horsepower needed to do JIT is pretty significant when you look at the actual computing power of an old ass computer.

Look at gigantic web operations like Facebook. They are running what you denigrate as "scripting languages" as one of the main components of an absolutely enormous operation. They are throwing money at the problem and speed is increasing.

Also, I believe that Java is a lot older than Ruby and Python. Both of which have gotten much, much faster in the past few years as companies are throwing money at tuning them.

There is nothing special about Java's speed increases, and the same increases can be had in damn near any language if attention is given where it is needed.

There's nothing special about Java. And Java is indeed the shining pinnacle of a language that's stuck in the 1990s.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

Indeed. Mind you Scala also runs on CLI. Kind of, support was not maintained between 1.4 and 2.8 and has just been restarted. But CLI based OS in depended GUIs are typical FOSS: 3 in-depended half backed solutions.

PS: I am a OSS (not necessarily FOSS as I sell GPL software for Symbian and Andoid) supporter but I know the weaknesses.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

Yes, indeed the article got it right. To bad the GP did not.

Yes, Scala is not a replacement. That was my point. To bad you did not notice without taqs.

Re:In a way that's reassuring.

[fishbowl]

I dare you to look that up on Monster/Dice :-)

Re:Shame Really...

[dumael]

http://www.microsoft.com/interop/cp/default.mspx

Shame really that someone couldn't even do the research to see if such wild claims about MS are in any way true.

Pointy haired boss.

[krischik]

And now explain that to the pointy haired boss.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Bigjeff5]

That's interesting, the spread for Lua JIT was greater than Java, but Lua had a slightly faster median than Java.

C++, contrary to what all the Java buffs say around here, was significantly faster than either of them. C++ median score was twice as fast as Java's.

Just goes to show you, if you want speed stick with native code.

It's too bad none of the .Net stuff is in there, they have C# mono, and it did relatively poorly, but mono could be significantly slower than .Net.

Re:Some insight from one of the bigger customers..

[fishbowl]

>Local programmers in both London and New York gotta eat too you know...

From the point of view of London or New York, if you can't afford to live there you can go to Dartmoor or Indiana.

There are people who see enough value in living in New York for a while that they will go there *at a loss*.
There are enough of these people that the labor and real estate markets are really competitive. It's tough to have to compete with people who already have enough money that they can burn it while working at a subsistence wage, but what can you do?

There are certain places I would like to relocate to (not NYC), but I can see that I have to compete against people who will take a lot less compensation that I could, just for the privilege of living in those places. And among those who do need to make a living, many have equity in the market there, and only need to cover incremental costs, where for me it would mean going into debt.

Re:Shame Really...

[Tubal-Cain]

ASIDE: Yes, you think it was clever calling it "see-sharp", but these are the same people who named their mobile operating system "wince", they are not that smart when it comes to basic language... 8-)

We'll need to rename GIMP before we can point and laugh.

Re:Some insight from one of the bigger customers..

[Raenex]

Personally, I am brushing up my C++, learning Objective-C and C#, as I think the medium and smaller companies in the market will start to migrate away from Java, as the cost savings of cheaper Java developers is lost once you have pay large amounts for the Java install and licensing.

So because Oracle announced an alternative, high-performance version of the JVM that you have to pay for, you're going to regress to C++, or lock yourself in to Apple or Microsoft? Brilliant.

Re:Some insight from one of the bigger customers..

[krischik]

One thing to remember - our bank gets and stays profitable because it pushes a lot of IT outside to third-parties (offshore developers are *much* cheaper than in London and NewYork)

Are your customers off-shore as well?
Or is cashing in unemployment benefit cheques all that profitable?

Java the next COBOL

[krischik]

In other words, Java will become the new COBOL

Funny. For some years I am lamenting that Java-programmers are the next COBOL-programmers: Inflexible and so set in there way that they are unable to learn a 2nd programming language.

Re:Java the next COBOL

[Cro Magnon]

I'm an old COBOL guy. Java IS my 2nd programming language, you insensitive clod!

The sky is not falling!

[Anonymous Coward]

To those of you that a). don't understand the Java and JVM options, b). claim to understand the Java market place and/or c). think the sky is somehow falling, please consider the following facts, all verifiable.

1. Oracle/BEA has for a long time offered a "premium" (paid for) JVM. It's called JRockit (with deterministic garbage collection).

2. Sun also offerred a "premium" (paid for) JVM, mainly for high-end support, but seem to have little penetration (probably evidenced by the fact that Sun no longer exists)

3. Oracle/BEA/Sun also offered "free" versions of their JVM. Oracle/BEA called it JRockit (without deterministic GC) and Sun called it Hotspot (that everyone uses and is now based on the OpenJDK since Sun open sourced it). You can still download these!

4. Oracle has simply confirmed the existing strategy - that's been an option for years. ie: You can still purchase a "premium" JVM that contains a bunch of stuff for seriously "high-end" customers (this will now combine JRockit and Hotspot high-end features - no point having two teams competing internally on the same ideas. Pretty standard stuff for Oracle when you consider the strategy they implemented with WLS from BEA). For everyone else, that use the OpenJDK, everything stays the same. That is, you're not forced to purchase a JVM.

Obviously it's much easier to jump to conclusions and pretend the sky is falling. Take some time to do some research before hitting the panic button. Might reduce your blood pressure.

Re:The sky is not falling!

[yuhong]

Yep, it seems that there are many commenters here who did not research the history of JRockit and HotSpot before commenting.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[ADRA]

No, forget for a moment that Java is faster than said scripting languages. Do these languages offer something more compelling than Java is the real question for me. If performance was the end all issue then fine, don't use a purely scripting language. But what I really want to ask is if Perl, python, Lua, etc.. offer anything to the developer that the Java language doesn't support? At least the logical jump from C/C++ to java was evident. It was basically the same language make-up without the worry of managing pointers directly, and having a large surface of standard language features built-in.

multiparadigm

[krischik]

Scala is multiparadigm and even when you ignore/don't use the functional part Scala is still a better language then Java.

Apart from that: What I really wanted to push is awareness of the differentiation between Java and JVM or Java-SE / Java-EE.

The original article got it right: this is an issue about then JVM and not Java.

Aesop explained this one

[FoolishOwl]

Well, as a manager, that's his job - to maximise the wealth of the company shareholders. If Java isn't making money (directly or indirectly), then he needs to institute change so that it *does* make money one way or another. Otherwise he wouldn't be doing his job (and then he'd be fired and replaced by the board of directors).

THE MAN AND THE GOLDEN EGGS

In other news, experts suggest that eliminating health benefits and blocking single-payer healthcare may lead to decreased worker productivity, that low taxation rates may lead to infrastructure decay, and that destroying the biosphere may have a negative impact on the long-term profitability of energy companies.

Hammer and nail

[benjymouse]

The best way to avoid hitting yourself in the head with a hammer is by not picking up the hammer in the first place.

Right. It is much better to ram in the nails with your forehead instead.

Re:Shame Really...

[vux984]

(*) In order it to be a "sharp" the symbol in use must be (1) in italics, and (2) in a musical clef. In Microsoft's language definition it is neither, that makes those two vertical and two horizontal lines a "pound" no matter how much they want you to call it a "sharp".

a) The pound symbol is that cursive L shaped glyph with the verital cross through it.

b) Only in the US is # called the 'pound sign'. Canada calls it the 'number sign', most of the rest of the english speaking world calls it the hash.

c) Technically you are correct that # isn't an actual sharp sign, but you are incorrect on both counts as to why. A sharp does not need to be on a "musical clef". And it has nothing to do with italics. The sharp must have true vertical bars, and slanted horizontal bars. A number sign must have true horizontal bars, with optionally slanted vertical bars.

d) The language C# is called C-sharp. Wandering around calling it c-pound and actually arguing that this is somehow correct is just pointless. Why "c-pound" and not "c-hash" or even "c-octothorpe"?

C-sharp is the clearly stated intention of the people who named it, and at the end of the day language rules are descriptive not prescriptive. The symbols use to write things do not dictate how we pronounce them. Written language is simply an approximation using a mix of tradition, convention, and convenience.

The programming language was named "c-sharp". It was then rendered conveniently as C#. Suck it up.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[Alex Belits]

building apps

Shut up, poser!

Re:Shame Really...

[rtfa-troll]

Reading even the second part of the first sentence of that, which I shall quote, since you are obviously too lazy to do so,

to the extent it conforms to one of the Covered Specifications, and is compliant with all of the required parts of the mandatory provisions of that specification ("Covered Implementation"), subject to the following.....

Nothing is "compliant with all of the required parts" of any standard. This is before we get onto all the exclusions. This document reads like it comes from a company that has been so used to getting away with evil that it has forgotten even how to lie properly. Trying to quote it to show Microsoft is okay is like trying to say that Ghengis Khan "traveled to many places" to show that he was a culturally tolerant kind of guy.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Windwraith]

I disagree. I code in Lua extensively and never found it behind Python or Perl in terms of speed. I'd even say it's actually faster.
Never did any serious benchmarks, but doing a procedurally generated world using Lua code, with a list of materials and complex items to populate dungeons, it's really really fast with Lua. Same with AI code and "actor states", control handling and projectile control. I have this program that runs 64 AIs, every one in it's own sandbox, one Lua state at a time (reportedly the "sub-optimal" way to handle Lua) and I get little to no latency every frame. Whereas python chokes under similar code (not bashing python at all, I love it and works much better for desktop apps). Oh my computer is terrible by the way, no modern game runs on it, but my own stuff filled with sub-optimal Lua runs pretty well.

Re:Some insight from one of the bigger customers..

[characterZer0]

Maybe by "offshore" he means off Manhattan to Brooklyn.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[gbjbaanb]

Oh I dunno, but think for a moment why half the web servers in the world run apps written in a scripting language. Answer that question, you answer your own - there must be something in it, or they'd all be running Java/C#/ASP. Wouldn't they?

Re:Sid you mean Java or Java-VM or Java-SE or Java

[Samah]

Yes, Scala has full access to Java APIs including Swing.

And if you wanted to remove one of the javax dependencies, you could use SWT instead of Swing.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[geminidomino]

(Also, what the hell is a 'scripting' language?)

Old-school vernacular expression for "interpreted" languages.

Re:Wow...

[master_p]

So, Microsoft, who tried to kill c++ with MFC in favor of Visual Basic will acquire Qt, the best c++ library? what an irony.

What about the Windows XP effect?

[TheNarrator]

I wonder if most Java development will stay on JDK1.6 and not move anywhere. That's sort of the problem Microsoft has with Windows XP.

Re:Shame Really...

[Kjella]

You're arguing with this guy ;)

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Ruke]

Oh, I have no doubt that Lua runs much faster than Python or Perl. That wasn't the comparison I was drawing. Lua runs much slower than C, C++, Java... compiled system languages are going to be much faster than scripting languages in almost every case.

Here's a nice blog entry visually benchmarking various programming languages that was posted on Slashdot maybe a year ago. It might be a bit dated; doubtless we've seen more improvements in the past year to relatively new programming languages, compared to those which have been around for decades. Still, it is pretty immediately apparent that with scripting languages, you're often trading performance for much less verbose code.

Man, revisiting that article makes me want to pick up Haskell. Everything I've seen makes it out to be a wonder-language.

Possible Patent Play

[darkchubs]

Ellison is out of his fu*king mind if he thinks the community wont build a better JVM. Obviously, Ellison is in control of his faculties, he's quite shrewd and I suspect, he intends to use the idea patents acquired with SUNW to sue anyone that makes the Open JVM competitive with their premium offering.

Re:Wow...

[shutdown -p now]

The new Nokia CEO is Stephen Elop, the former head of Microsoft Corp.’s business unit.

If you're going to look at past work experience of all senior officers in all major corporations from that perspective, the only conclusion would be that everyone tries to assimilate everyone.

I mean, you don't think that MS execs are secretly grown in vats, in adult form, in an unnumbered building on 1 Microsoft Way, Redmond, WA - right?

Re:Sid you mean Java or Java-VM or Java-SE or Java

[jpate]

the real question is: Can it web scale?

probably.

Re:Some insight from one of the bigger customers..

[raboofje]

Stallman wrote the Java trap, and we all laughed. Sun is nice we thought, it'll be ok. We were all wrong. Stallman saw further, he saw that even if Sun was ok, if someone bought Sun, then things could get messy. Welcome to messy.

Actually, most of the platform has been open-sourced by Sun before the takeover under the 'OpenJDK' name.

Indeed, GNU now carries a 'headnote' to the article explaining the new situation: http://www.gnu.org/philosophy/java-trap.html

The situation is still messy - but not for the original reasons laid out in that article. If worst came to worst, the JVM can be forked - but forking a widely-used platform is a whole other can of worms, of course.

Re:Shame Really...

[creepynut]

I'm pretty sure the reason it's called C-sharp is in reference to the sharp symbol used in music.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[jpate]

First of all, a higher education courses which are based on programming languages should be designed to educate the students on fundamental topics such as programming paradigms and design patterns. It is unthinkable for a higher education course to waste both the teacher's and student's time doing vocational training. After all, if you are enrolled in a higher education program then you should be smart enough to handle that stuff by your own. Llet the teachers teach the important technical and scientific concepts which are harder to grasp.

What? First, you say the focus of a programming course should be on the abstract side of programming. Sure, sounds good. But then you say:

Then there is the reason of applicability. If a course is based on a technology which doesn't have a stable basis and which is likely to drastically change due to a project leader's whims then it is certain that the code written in that particular technology will suffer from bit rot and that your skills which you earned throughout your course will essentially become meaningless. In short, due to a technology's dynamic nature, those who invest their time and effort in a degree or course which is based on that single tool will soon find out that they have wasted their time.

Then there is the issue of control. If a language is not standardized then you don't have any assurance that your code will run at all in the future. This issue is particularly problematic if a language does not provide any specification regarding the implementation or is based on a single reference implementation whose project leadership often invests their time and energy in complete rewrites.

So, knowing this, languages such as Python, Perl, Lua, Ruby or whatever trendy language you just picked up is not nor it ever could be sanely adopted by a higher education degree. Doing so is a recipe for disaster and it does a disservice to all students.

So a course that doesn't give you vocational skills will have been a waste of time?

Re:Shame Really...

[exomondo]

(*) In order it to be a "sharp" the symbol in use must be (1) in italics, and (2) in a musical clef. In Microsoft's language definition it is neither, that makes those two vertical and two horizontal lines a "pound" no matter how much they want you to call it a "sharp".

And OSX should be O.S.X., iPhone should be Eye-Phone, Windows XP should be Windows X.P., etc... It's an approximation, the vast majority have no problem dealing with it, im sure if you apply yourself you can eventually get your head around it too.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[zhong-guo-1]

I really thought that I can't believe you guys didn't see the tongue in cheek here. I thought I made it plainly clear that I was mocking a lamer

Re:Opensource and corporate control bad combinatio

[yuhong]

Another IMO worse example is MySQL AB themselves, before they got bought by Sun. They were going to do a support model, but the big problem is that the distributions already is taking all the support money. So they tried to make money by providing support on Windows. Obviously this was flawed, so eventually they tried to make some new features proprietary (in a different way than what Oracle is doing with Java), resulting in a backlash. MySQL was almost going to IPO with such a bad business model when they got bought by Sun.

Re:It's platform, stupid ! - what about parrot?

[antifoidulus]

Actually Parrot(and Dalvik for that matter) are much better VMs than the JVM in theory because unlike the JVM, which is stack based, they are register-based which means its MUCH easier to implement the VM, or at least parts of it, in hardware. The JVMs stack based VM was a neat idea, but it makes hardware acceleration of the JVM much more difficult, though certainly not impossible(people have implemented Java-on-a-chip for instance). With the popularity of embedded devices recently, a VM appropriate for them as well as the desktop makes a lot of sense, but thus far there hasn't been a lot of interest in running Dalvik on the desktop.

Re:Shame Really...

[vux984]

I'm pretty sure the reason it's called C-sharp is in reference

Of course the name C-sharp, and the "C#" spelling is -intended- to invoke the musical C-sharp.

The reason for the symbol choice itself is simple.

You start with C.
The subsequent iteration is C++ (using the C increment operator).

So the next iteration is C++ ++
Then cleverly rearage the two ++ on top each other:
++
++

and stick all them together...
#

and call it C-sharp because both C-pound or C-octothorpe sound stupid, while c-sharp is clever.

Money money money

[sashang]

Now all Java devs will be able to afford a yacht like Larry's after customers fork out for the non-crippleware version of the jvm.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[Qubit]

PS: I am a OSS (not necessarily FOSS as I sell GPL software for Symbian and Andoid) supporter but I know the weaknesses.

There's a common misconception that Free Software (capital "F") equates with "you can't make money on it" or "you can't sell it." Nothing is further from the case.

There's nothing wrong with selling Free Software. In fact, the Free Software Foundation points out in their GPL FAQ that "The right to sell copies is part of the definition of free software. Except in one special situation, there is no limit on what price you can charge. (The one exception is the required written offer to provide source code that must accompany binary-only release.)"

So feel free to develop GPLed software and sell it. It's all good.

Re:Shame Really...

[Qubit]

because both C-pound or C-octothorpe sound stupid, while c-sharp is clever

What, the name Coctothorpe doesn't sound awesome to you?

I think it's almost as strong a name as Octopussy...

...now wait a second. What if we named it Octothorpussy? We could shorten that to #Pussy or just #P, and then just colloquially call it "Sharpie".

Re:Sid you mean Java or Java-VM or Java-SE or Java

[krischik]

I already sell under the GPL. About one in 100 customers actually request the source-code and I am more then happy to send them the additional download link. Since my customers use GSM networks for download binary-only-releases are in there best interest.

I just thought that the FOSS being Free and OpenSource might be stricter then the GPL itself. Thanks for the clarification.

Re:java dies....

[smash]

"troll". slashdot for "i disagree but am too inept to formulate a coherent response".

Re:Sid you mean Java or Java-VM or Java-SE or Java

[theshowmecanuck]

Your astroturfing for Scala all over the comments for this article is absolutely fucking annoying.

Re:java dies....

[Ash-Fox]

A neat idea, back in the mid 90s of "write once, run anywhere" that ended up so horribly broken that it became "write once, debug/rewrite everywhere" only with shittier performance than native, recompiled code.

[Citation needed]

Re:Java is Dead, Write Once, Debug Many Times

[Ash-Fox]

Java is now officially dead.

A free java, a paid java, and 100 million java clones out there.

It's time to take garbage collection out with the trash.

Write Once, compile on 20 different java interpreters, and cross your fingers and hope it works.

As long as I can remember, it's always been like that with JRockit and HotSpot. How is it dead now?

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Ash-Fox]

Oh I dunno, but think for a moment why half the web servers in the world run apps written in a scripting language.

That's easy, live editing of programs without compilation. Interesting to note, many large scale systems are using asp, jsp etc. and even scripting languages that end up in large scale systems, end up with special "cached" pre-compiled code on systems that deal with scaling up.

In reality, from the way I understand it, it's just from ease of use to program live/edit rather than the actual language it self. But I don't think this answers at all what the grand parent wanted to know, so... I don't really get your point either.

Re:Shame Really...

[scooterjohnson]

You made my day, sir. If only I had some mods! Instead, I guess I'll have to name my firstborn "Coctothorpe".

Very true

[krischik]

Very true. Of course I was well aware that just changing the Language Java is no the solution. I guess some sarcasm tags are missing from my postings.

Re:Uh, watever, just migrate to Python, Perl6, Lua

[Windwraith]

Oh, my bad! Yes of course Lua is slower than C, that's undeniable.
I just develop games, to be honest, didn't really find interest on "the desktop". C/C++ + Lua is a really speedy combo allowing for great realtime stuff.
I heard a lot of Haskell evangelism too...dunno if it's fanboyism or fact, though.

Re:Sid you mean Java or Java-VM or Java-SE or Java

[mcvos]

If you want to know what's being discussed, just read the summary. It's the JVM. And Scala uses the very same JVM. So not much of a replacement.

How is this a troll?

[IBitOBear]

Seriously, Patent issues abound. How is mentioning them a troll?

Or do you object to the name C-pound?

Re:Shame Really...

[IBitOBear]

The _English_ pound, as in pounds-sterling, as in money, is the cursive ell with a bar.

The pound, as in weight, as in 16 ounces, is a cross-hatch.

That makes "#" a pound.

It could also be number sign.

Technically it's an octothorp used to designate numbers or weights.

The language is c-pound. Or c-octothorp, or c-numbersign.

I can explicitly say that I wrote down the Letter "C" followed by a cursive ell with a line through it with the express purpose of naming a language
"see fog" but the symbol would persist in being an english pound sterling.

The problem is that nowhere is "#" a "sharp" except in musical notation.

It's see-pound... 8-)

(yes, this is flamebait... but the language is burning rubbish full of patent bombs. Sometimes it _takes_ flamebait to make people see the fire ready to burn them to death. 8-)

Re:Shame Really...

[IBitOBear]

I am, and was, fully aware of the usage and intent that "C-#" was supposed to be "see sharp" as in music.

If you examine a sharp sign, it is two horizontal lines and two vertical lines, the _second_ vertical line is _higher_ than the first, in indication that the tone is supposed to be higher than the actual note by a half-tone. The "sharp" indication, which really only exists in the musical clef, is _distinct_ from this symbol "#".

In the octothorp, there are two horizontal lines, and two vertical lines. The vertical lines are the same height and drawn from the same baseline. They are usually slanted such that there are no right angles.

In no case have you ever seen "C#" written as "see sharp". That leaves "see number-sign, see-octothorpe" or "see-pound".

There is a unicode symbol for "sharp", it is \u266f, as opposed to 0x23, the "pound sign".

Seriously, there is a _huge_ difference in symbolism.

It's not _my_ fault that the same people who named a mobile platform "wince", named a language "see pound" while failing to be clever.

http://www.fileformat.info/info/unicode/char/23/index.htm

http://www.fileformat.info/info/unicode/char/266f/index.htm

Intentions only take you so far when you make an epic screw-up. They named it something, but it wasn't "see sharp". 8-)

Re:Is it the "see pound" thing?

[IBitOBear]

If you think I am wrong about "C#" being read "see pound" please check the unicode symbol definitions:

A pound/octothorp/number sign has horizontal lines a two (usually slanting) vertical lines drawn from the same baseline:

http://www.fileformat.info/info/unicode/char/23/index.htm

vs.

A musical sharp sign has two vertical lines, the second drawn higher from the baseline than the first, the "horizontal" lines incline. All of this is intended to indicate the half-tone rise of the notes so designated.

http://www.fileformat.info/info/unicode/char/266f/index.htm

You will node the "Index Entries" in both.

The name is "see pound"... sorry, just a fact.

Technical citations

[IBitOBear]

Pay close attention to the symbol names in the index entries part of the table.

The musical sharp sign:

http://www.fileformat.info/info/unicode/char/266f/index.htm

The "horizontal" lines rise to the right to indicate the rise of tone. The second vertical line is drawn to a higher baseline, also to indicate the rising tone.

Now...

The pound sign, aka the octothorpe or number sign:

http://www.fileformat.info/info/unicode/char/23/index.htm

The two "vertical" lines are usually rendered leaning right (but pure vertical is okay). The horizontal lines are truely horizontal and are drawn to the same right-hand and left-hand bounds.

Now: "#" that isn't a "sharp", its a "pound", or number sign, or octothorp or cross hatch. Period.

So this: "C#" is "see pound"...

Sorry, its just the facts man.

(And yes, its a sad fact that looks like a troll or a flame-bait, but that isn't technically my fault. Talk to the "wince" people... 8-)

Re:Technical citations

[vux984]

Pay close attention to the symbol names in the index entries part of the table.

Why? I know all that, and you are repeating what I already said. I'm well aware # is not the correct sharp sign. I didn't argue it was.

I argued that it simply didn't matter.

So this: "C#" is "see pound"...

No. Go read any music discussion on the web...here's just one excerpt of millions...

For example, the notes in D Major are:

D E F# G A B C# D

While the notes are the same, you would never say,

D E Gb G A B Db D

They have no problem using # for sharps. And as it happens they also use the lowercase b for flats.

Written language is to convey the intended meaning. When the intended meaning is "sharp", then its "sharp" whether they write the word "sharp", use the "#", or use the proper Unicode sharp symbol.

Re:Technical citations

[IBitOBear]

Lack of choice in ascii for a sharp is sad. And Even here in SlashDot I cannot post the unicode character for sharp.

But in no case does that mean that Microsoft named the language "C-sharp" when they put the letter C and followed it by a Pound Sign.

In _their_ materials they had the full choice to use an actual sharp. heck, they got the little "tm" trademark symbol in there, and they used the proper circle-c instead of the legally junct "(c)" construct.

So what they did was name the language "see pound" on purpose, or due to incompetence, because they were trying to be "clever" instead of "clear".

The fact that there are limits to the ascii character set that people evade typographically is completely beside the point. Microsoft wasn't limited to the ascii character set when they decided to call the language C-pound. Windows had unicode support by default by then.

They just did it because they are dumb.

It was an _utter_ fail on their part.

Apologize for it all you like. Justify it all you like. It's still C-pound.

And yes, I will accept C-numbersign and C-octothorpe as valid aliases, since both are also correct... C-pound is shorter and more obviously in line with the awful language constructs in use, so it is my preferred.

It is, however, by no reading, a "sharp" and it was pretty blunt of them to try to make it one. 8-)

Re:Shame Really...

[IBitOBear]

Eh, nothing I can do about that, and at least "GIMP" means something: "GIMP is the GNU Image Manipulation Program..."

What is "C pound" supposed to mean? 8-)

Reading the Article

[krischik]

I did read the article. Slackware95 did not. Slackware95 wanted to replace Java while the JVM is on stake.

Re:Shame Really...

[vux984]

In no case have you ever seen "C#" written as "see sharp".

"In no case..."??

Of course I have. Thousands of times. Did you even look to see if what you are writing is true? Look at pretty much any musical discussion anywhere on the internet...

http://www.list-of-chords.com/

http://musiced.about.com/od/lessonsandtips/qt/minosclaes.htm

http://www.guitar-chords.org.uk/chords-key-b-flat-minor.html

http://www.ehow.com/how_5496374_play-adopted-cyclopedia-music-theory.html

http://www.8notes.com/guitar_chord_chart/ax.asp

Seriously, there is a _huge_ difference in symbolism.

And yet musicians the world over have accepted it as a suitable convenient ascii substitute for sharps to the point that # and b are accepted conventions when discussing music on the web.

Intentions only take you so far when you make an epic screw-up.

The only epic screw up here is insisting that it be called C-pound.

Re:Shame Really...

[IBitOBear]

It's "C pound", live with it... 8-)

Just like (c) isn't the circle-c for copyright and it has no legal meaning no matter how many people put it in their code, this thing ==> # == is not a sharp. Ever.

To be doubly firm

[IBitOBear]

Does that mean that the string "a b c" is "a sharp c" because I said so?

No. Because that isn't the meaning of the symbol "b". It is _necessary_ to misuse "b" in a musical context if you are stuck with ASCII encoding.

Microsoft was not so stuck. They either didn't know, or didn't care, that they were naming the language C-pound.

The license to pretend that some things are actually something else has its limits. The numeral three isn't the letter "e" no matter how many people write "l33t".

Deep in your soul you know I am right, else you would have stopped two messages ago... you are a C-pounder. And every time you preface a function with square-bracketed config directive looking thingies, your soul dies just a little more. Pounding it deeper into ignominy.

Re:To be doubly firm

[vux984]

Deep in your soul you know I am right, else you would have stopped two messages ago...

Right back at ya. You can't let go can you?

Re:To be doubly firm

[IBitOBear]

Why should I? We both know I am right. It's C-pound, or C-numbersign, or C-octothorpe. In no case is it C-sharp.

I give technical citations, you give excuses.

Unless you claim that I can write "A b Tire" and have it mean "a flat tire" because ascii doesn't have a the musical "flat" so I had no possibility of clarity...

Re:To be doubly firm

[IBitOBear]

Your OSX analogy is false, and a straw man to boot.

I don't insist that it has to be Oh Ess Ecks because the upper case X _IS_ the correct symbol for the roman numeral representing the value ten. (it isn't "10" by the way, it's ten, but that's beside the point.)

So there is nothing wrong with either statement regarding "OSX" and calling it "ten" or "Ecks"

But Dude, "#" is not and never was a sharp, it's a pound, or a number sign, or octothorpe. A sharp is, and always has been _completely_ dissimilar to a pound sign. In size, sharp is smaller; in baseline sharp is higher, in slant sharp is upright and lines that rise to the right.

If you argue that "#" is "sharp" then cleary "Q" and "O" are the same glyph, numeral one (1) and lower case ell (l) are identical, and "3" and "e" are interchangeable. But they aren't.

"All their documentation" calling the pound in C-pound a sharp don't make it so.

They botched.

And you have swallowed that botch like a good little camper. 8-)

And yes, now I am just greifing you, accuracy and correctness are clearly not things you care about.

Re:To be doubly firm

[vux984]

We both know I am right.

Personally I doubt anyone thinks you are right, including you.

Unless you claim that I can write "A b Tire" and have it mean "a flat tire" because ascii doesn't have a the musical "flat" so I had no possibility of clarity...

If you wrote Ab to mean A-flat, that would be fine. You're example is just absurd.

In the case of c-sharp, they named it c-sharp to play on the fact that a C-sharp is a step above C. They are explicitly referencing the musical semantics. And then they wrote C#; within the constraints of ASCII.

I do not see that as anything approaching the absurd example you tried to constrive involving flat tires.