Is Java as we know it doomed?

While Oracle and Sun Microsystems are hailing Oracle's purchase of Sun [1] as a big boost for Java, others are not so sure, questioning what kind of control Oracle might try to exercise over the popular platform that has driven so many enterprise applications since it was first developed in 1995. Observers also expect Oracle to make a go of trying to make more money off of Java than Sun ever could.

Sun has tried to leverage Java as a lead-in to selling services, but without much success. By contrast, Oracle is very disciplined about extracting money from its technologies.

"Java is one of the computer industry's best-known brands and most widely deployed technologies, and it is the most important software Oracle has ever acquired," the companies said in a joint statement announcing the acquisition. "Oracle Fusion middleware, Oracle's fastest growing business, is built on top of Sun's Java language and software. Oracle can now ensure continued innovation and investment in Java technology for the benefit of customers and the Java community."

[ What on earth was Oracle thinking when it bought Sun? InfoWorld's Neil McAllister shows what the real motivation may be [2]. | Sun's efforts to make Java open source have also led to decidedly mixed reviews [3]. ]

Oracle's commitment to invest in Java may mean it will go full-force at trying to make money off of Java, a path Sun has not pursued strongly. "I think that they will realize that they want as many people using Java as possible so that's good for their middleware," said Rod Johnson, CEO of SpringSource and developer of the popular open source Spring Framework for Java application development. The acquisition was partly defensive because Oracle probably did not want competitor IBM owning the Java language, he added.

Oracle has been "better at making money off everything than Sun was before," concurred Vivek Ranadive, CEO of Tibco, a middleware company that competes with some Oracle offerings. "They're not beyond finding ways to make more money off something, and it's always at the customers' expense." When buying companies previously, Oracle has sought to increase maintenance revenues, he said.

Does Java's open source status help it? Theoretically, Sun is not supposed to be the owner of Java: It has offered up its version of Java to open source. The Java Community Process (JCP) has been set up as a multiparty organization to amend the platform. But Sun has remained the dominant force in Java, with the company always at the forefront of improving it.

The open-sourcing of Java has put the technology out into the community at large, and that is where innovation in the platform now comes from, Johnson stressed. "The innovation in Java comes largely from open source," Johnson said Monday, reaffirming similar remarks he had made last month when it was IBM, and not Oracle, that was supposed to buy Sun [4]. "The language itself is open-sourced. I don't really see [Java] as something that Oracle can own in a meaningful sense."

Less optimistic, though, is Tibco's Ranadive, who asked whether Oracle can be trusted not to manipulate Java to its own ends. Oracle competitors such as SAP rely on Java, and they now must consider the impacts of the Sun acquisition, he said: "As you continue to put more eggs in the Java basket and your biggest competitor owns Java, what you do?"

And Ranadive doesn't think Java's open source status means all that much. In Java's case, the label "open source is a bit of an oxymoron. It's really not open as such," he says. "All control [of an open source software project] rests with the party that offers it. It used to be Sun and now it'll be Oracle." Ranadive anticipates Oracle will dominate the JCP.

SpringSource's Johnson also expressed fears over Oracle's effects on the JCP. He serves as a JCP executive committee member. "I think that there will be a lot of concern about Oracle potentially making the rules for Java," he said. Ironically, that fear may force the open source community to more aggressively assert stewardship of Java, he added.

Neil McAllister, an InfoWorld blogger who speculated earlier this month that Oracle might buy Sun [5], anticipates changes in the JCP. "I think you will see Oracle having a lot more heft in the JCP but also maybe it will change direction a little bit, as far as what areas of Java development it sees as being important or relevant," he said.

How Oracle's ownership may help Java At the Eclipse Foundation, an open source tools organization that has counted Oracle as a member but not Sun, Executive Director Mike Milinkovich sees the merger as a "very positive sign for Java and open source." Oracle "will be able to provide the resources and leadership to continue the innovation in the Java community. I see their support for OSGi [6] and Eclipse Equinox [7] as being key to driving the next generation of runtime middleware based on the OSGi standard," he added.

Also optimistic is Matt Asay, vice president of development at Alfresco, who said Oracle will be good for Java. "In some ways, it might even be better than Sun." That's because Sun has "never been willing to fully let go of the reins. Oracle doesn't need to monetize it directly. It just wants to make sure that Java flourishes," he said.

Battles with IBM foreseen Bill Roth, vice president of product management at GSI Commerce and a former Sun and BEA employee who left BEA after Oracle acquired the company last year, was not optimistic about Java's fate. "I believe that this acquisition means the death of Java," he said."I believe fighting between IBM and Oracle will lead to the end of 'write once, run anywhere'" for Java, Roth said. "There will inevitably be some disagreement between the owner of Java, Oracle, and IBM, [which] has invested billions in the technology. I cannot imagine that IBM would blithely let Oracle determine the future of a technology so embedded in its software stack without a fight. IBM's only recourse will be to fork the code," he added.

Microsoft, Adobe may benefit A main benefactor of Oracle buying Sun is Microsoft, said Tibco's Ranadive. Companies that have had a mixture of Java and Microsoft .Net software might opt for Microsoft now, with questions arising over whether Oracle will keep Java open, he noted.

Oracle also might de-emphasize Sun's JavaFX rich media platform, McAllister said, which could help both Adobe and Microsoft, which offer the competing Silverlight and Flash/Flex technologies. Neither Silverlight nor JavaFX have made huge inroads [8] among developers.

In what cases does java execute faster than C++? Cuz I aint buying it. Not by a long shot.

This is from the wikipedia article on Java vs C++ performance. Also check references 7 and 8.

Oracle kills MySQL and acquires a hardware/server division. Java is incidental.

Or Oracle makes MySQL more compatible with Oracle DB, and puts a ceiling on MySQL performance so that people migrate to Oracle DB as their performance needs increase.

That’s basically what I meant. MySQL as a free alternative to Oracle DB and comparable in performance wont exist. I wonder if it would be still “legal” to run earlier versions of free MySQL in production environments.

Aside from C/C++/C#, I also work with python for scripted tasks. Functional programming caught my attention via the Erlang language. It is a good choice for working on a multi-core machine with shared cache. Once a variable is "bound" with a value, it becomes immutable. That is important to ensure that other instances operate consistently when using the variable in a decision making process. Ericsson uses the language for telephone call processing switches.

I work on a project that started out as pure C++. The source code base was under 5 MB as pulled from an SVN repository. There was a cool moving map (tied to GPS) display that was desired by certain persons on the project. The code was written in Java. That became an excuse to reimplement the display as a Java application. Inside a couple of weeks, the SVN extract into the sandbox was over 450 MB. We had code that was a mix of Java/C++. It was necessary to bring in Eclipse (another 500 MB of disk space) and SWIG to build the damn thing.

The Java of 2009 is fatter and uglier than what I left behind in 2000. It's certainly the last thing I would put on an embedded system. Conversely, the gcc/g++ compilers continue to get better. The code is getting smaller and faster. Exactly what I want in an embedded system.

Java is not best suited for graphics apps. The AWT can bog down a system and cause poor performance. Java’s forte is *server side computing* - business logic, security, database connectivity, messaging, remote computing, web apps..etc. As for embedded systems, I have no experience with them but I know Google Android is Java based. Google even publishes an API so you can write your own apps for your phone.The Blu Ray specification(BD-J) also calls for a Java implementation.

I have a couple Sony BluRay players. They work well, but the Java aspect causes start up delays for 1 to 3 minutes on titles with special Java features.

In 1996, I was using Java for server side CORBA. The JVM garbage collector made amends for the crappy state of CORBA code at the time. The C++ versions leaked memory. Java did too, but had a means to clean up. In 1998, I joined a project in progress. It was a giant applet with Oracle stored procedures for the back-end data. It performed terribly on the Netscape browser. I found a nice profiling tool and discovered most of the problem to be handling of memory by the garbage collector. I ported the applet to run on Microsoft's JVM. It ran MUCH better. The MS JVM had a superior garbage collector. I further improved the code by eliminating thousands of "font" objects and replacing them with a fixed set of about 20 common font objects. That reduced the memory footprint and reduced startup time. Just as the new, fast implementation was ready for delivery to the customer...Microsoft abandoned their JVM. We were stuck with Netscape/Sun again. Barf.

I read through it - and I must admit - I was surprised. I had a few thoughts I felt like sharing after reading -

While it may be true that C++ is harder to optimize - I don’t see that as a valid point in a speed comparison. The proper use of optimizers and debuggers go a long way to dispelling those issues.

For C++ on a Windows/Linux platform, you should really go with Intel’s compiler. In my experience, it produces the fastest code. Comparing to free/legacy compilers just aint right.

I understand in theory the benefits Java could take advantage of in coming versions - but properly written C++ code can as well.

For anyone who has programmed both - especially for a Windows/Mac environment, or as a Web Service - the reality of the situation is that C++ outperforms Java, has less issues, and requires less overhead to maintain.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.