Is Java dead?

No Java is certainly not dead and is more alive than Ruby, Scala, Groovy, or any of the languages that run on the JVM. The only people who think Java is dead are the vocal minority that happen to like languages that nobody uses or cares about. If any of those languages catch on, they will declare that language dead and move onto the next obscure language. Meanwhile, Java will continue to evolve and will remain king.

I think C# should have been in your list of successors to Java.

Now that I actually read some of your article I see that you say that Java is actually not dead, is not fundamentally flawed, that there is still a lot of interest in it, etc. I agree that it is healthy to explore other languages, I just hate it when people assume that their favorite language or web framework is better than Java then declare that Java is dead. There are still millions of Java developers that actually like Java, want to see Java continue to evolve, and will continue to use it as the default choice on new projects for startups and enterprise use.

Why can’t we all just get along? I don’t like Ruby but I don’t write blogs about how Ruby is so 2006 or that it has been forgotten now that Scala and other shiny new languages have emerged. I have lost a lot of respect for Ruby programmers because they are so completely hostile towards Java and anyone who doesn’t like Ruby.

The next language I’m going to learn is JavaFX Script.

I don’t wont get into the the debate if Java is dead or not. I just wanted to add that mint.com uses Java in its backend as does last.fm (at least to my knowledge).

While last.fm is a little older (founded 2002) both are huge and are profitable. If that seems too old, consider that LinkedIn was launched in May 2003.

@stephan: That’s besides the point though. For all practical purposes, C# is a much improved Java (properties, events, lambda, inference etc). If “migration” is your benchmark, you can not include an exotic and complex language like Scala in your list which only appeals to the top 5-10%. Also, if you notice, many critics of JDK7 make references to the usability/features of C#.

Java is dead in an innovation sense, Sun has made that very obvious and plan to make $ on JavaFX instead. So while Java certainly is not going anywhere… it’s also not going anywhere!

I think, that when SUN fails, google would continue the development of the JVM. Most of the best Java-Coders are working for google.
Last week they announced noop (http://code.google.com/p/noop/), so i dont think that Java would vanish in the next years.

You left checked exceptions off of your “bad ideas” list. Intentional?

-m

While I agree that Java is slowly dying, your numbers are greatly exaggerated.

When searching google for phrases like this, please add quotes: “Java is dead” actually only has 172.000 hits.

Hmm, but what about Java’s use on cell phones? While smartphones are getting more popular, they’re not reachable by everyone. Last I knew, lots of phones still used Java (and thankfully, few use Brew), mine included (though I hear it’s US counterpart uses Brew – first time I heard of it!).

“[Java] is obviously not any longer the default choice for web-sites and web startups. For some years this has been Rails, and with good reasons.”

Do you have any citations to support this? All the statistics I could find about programming language usage/popularity clearly and dramatically indicate that PHP has been the default choice, not Ruby. It’s not even close.

So, do you think C# would suffer the same fate as Java?

Should your scientific research via Google search have been “”java is dead”" as opposed to “java is dead”? or, potentially “”java is dead” OR “is java dead”" in which case you get hundreds of thousands, not millions of results.

at the end of the day, i would say know. most of my friends are java coders and they work in the financial sector, which is heavily dependant on enterprise java. if money makes the world go round it will make sure that it keeps making java go around.

Wow! This is the most thorough exploration of “Java is dead” that I have read, and your conclusions seem very reasonable to me. (FWIW, people at work have mostly been using Groovy [on the backend/JVM & frontend prototyping] and C# [on the frontend/ASP.Net] for new work.) The comparison between Java’s level of “life” and that of C++ will become closer and closer. What in turn will supplant Java? We don’t know yet.

Java is nowhere close to being dead, inasmuch as it is finding a space where it is likely to sustain itself for a very large period of time. I suspect in many ways as C sustains the plumbing for Systems Programming, Java will continue to sustain in many cases the plumbing for application programming especially for the stacks written on the JVM.

Is Java the language dead ? It will find a space for itself wherein it will continue to be strong and may even be irreplacable. However this space is unlikely to be the application programming space which currently occupies the market share of a lot of development activities and developers. What is clearly apparent is that it is losing mindshare fast, and consequently it could lose some amount of marketshare as well. However even if that were to happen, it would require a significant amount of time since it is far easier to change a technology than the installed base of developers competent with such a technology.

FWIW I had written a post on a similar topic (but covering some very different points) a year back Java : the perpetually undead language.

Anecdotally, I’ve seen many enterprise shops migrate internal app development to .NET from Java, so C# is a valid entry in the list.

[...] 21 September 2009 Posted by manniwood in Java. trackback Stephan Schmidt’s Is Java Dead? is a perceptive look at how Java is not dead, but how it is nonetheless not the only language one [...]

“…does Java no longer meet it’s goals?”

Its goals, not it’s goals.

Hi Stephan,

as Ricky suggested, C# should have been there – in my opinion. You are right when you say that the two are very different in terms of background philosophy/culture (open source vs. proprietary). However, both languages have very similar syntax (a Java coder can learn C# VERY fast) while the C# technology is much more innovative than what happens in the Java space right now, I think.

I am working with both on a regular basis for now almost 2 years. To put it short, both have features/weaknesses (Eclipse rocks, Visual Studio sucks, ASP.NET rocks, JSF sucks) and share most of the “enterprise application development” space, I think.

It is true that someone that is used to Java & open-source technologies (Linux, BSD, Apache, etc) will most likely not switch to super-proprietary Windows/IIS/.NET combo. But, as a programmer, I like to think that someone mastering both technologies won’t be afraid to choose a technology for its qualities rather than because he thinks “the other choice” sucks, without knowing much about it.

That being said, Java is definitely not dead. The innovation just won’t come from Sun or the JCP I think. They moving too slow, so third parties are taking over. Check out what SpringSource is doing, nice innovation. And, on the other side of the fence, Microsoft labs are developing .NET technologies very fast (compared to Java, that is).

Just my 2 cents :-).

Great post. IBM also has a very good, although a bit old, article on Java’s Death: “Dead like COBOL” (http://www.ibm.com/developerworks/java/library/j-cobol.html?S_TACT=105AGX02&S_CMP=EDU).

Interestingly enough, this is linked from Scala’s site (http://www.scala-lang.org/node/960).

Don’t expect Java to lose significant mind share for the next 5 years at least, let alone die. Only the enthusiasts (which is a minority) is worrying about Java’s death right now, and not without reason. The enterprises are happy running Java — when Java is replaced they will simply follow suit. It ultimately depends upon how well a new language delivers and gets adopted, which would eventually replace Java.

Java is not dead, what is dead is the idea that one system can solve all problems in the world.

for example, java has tons of mature middleware that can be certified in big corporate project.
rails does not, it has the bare minimum.

but rails allows to build medium size project an order of magnitude faster than java big guns.

there is no magic, you can drive faster with a small mini car in a city and you can carry more with a 16 wheeler.

i prefer several tools for different jobs.

Great article – thanks. Long-time user of Java, later .Net, back to Java for App Engine. Will check out Scala.

Typo here: “the early lock and synchronize system of Java proofed to be too difficult”

proofed -> proved.

Hey, and I also had an Amstrad CPC-464 and an Amiga 500 :) Still have the 500 in storage…

From a language perspective, C# should be included. From the JVM perspective, C# (and F#) doesn’t play (yes, some shops are migrating to other platforms). Is the JVM dead? No. Is a language dead if it cannot, or will not (BDFLs..), evolve due to it’s own legacy? Maybe. Maybe alive/dead binary thinking is not very useful here.

As for developer inertia, I run in to an awful lot of blub when I talk about Scala.

@stephan: i know mate, it’s called tongue in cheek ;)

point remains, when you narrow the query to “java is dead” or “is java dead” you get far fewer results than just searching for articles that contain the words “java” “is” and “dead” as opposed to the phrase.

I left this comment on another site that referenced your article (http://mpwoodward.posterous.com/):

Interesting commentary on the state of Java. I agree with many of your points and find the article to be a pretty balanced assessment on the state of Java. I do think you should have included Flex/Flash in the list of potential successors. I know it is not as popular with the JVM crowd as say JRuby, Scala, etc. but Flex is a solid way to develop both web and desktop applications. Especially if you put something like Java behind it. I particularly found the point about Java being an OO-only (Object Oriented) language interesting. Especially since it seems that since Java became the thing, any language that was not OO was considered inferior. However, there is a place for OO and there is a place for procedural coding. Not everything needs to be an object.

I believe Java as a language will be around for while, but I think that Sun really missed the boat when they made the web development process as complex as they did. I’m sure Java purist will point out the IDEs and all the different tools they have to help write much of the boiler plate code and speed up the process. But ultimately why should the process be some complex in the first place?

IMHO, Java still has a place as an enterprise application language or back-end language for someone looking to add strength to apps that require very heavy lifting behind the scenes may utilize Java, but I think that as the new languages evolve this issue will go away to the degree that Java may eventually become looked at as COBOL is looked at today, a useful language in some instances, but not what you start up a site on today.

Java, the language, in its current form, IS dying. Let’s be honest. The language needs an overhaul. A lot of smaller projects are making runs to try and make this happen in order to make the language a little bit more expressive. I see the language forking at some point in the future. Java, the language, needs to break backward compatibility and let the language mature. As it stands now, it’s too verbose and you have to say a lot to get things done. I was really disappointed by the lack of changes accepted for 7. I thought that the climate was perfect for core language changes to be introduced for 7.

I don’t think the JVM is dead.

[...] take a few things from this blog post on “Code Monkeyism” and discuss: “Searching for java is dead with Google, one [...]

I was about to say the chance of Java dying is the same that COBOL or ForTran have of dying (none!). Then I thought a bit better, and I’ll posit something else:

The chance of Java dieing is the same chance C has of dying.

And that is something else altogether. While C still has a huge place, C++ and Objective C have eroded its space much faster than I’d have believed. Still, it won’t cease to be important to a lot of people, as the same applies to Java.

But one can ask a completely different question, which people will often equate with “Is Java dying?”. That question is:

Is Java losing its proeminence?

I won’t answer that question directly. I’ll point, instead, to something Paul Graham (I think) once mentioned. First, people identify themselves with groups, institutions, beliefs, etc. By “identify” I mean the people will see attacks on these things as personal attacks.

Second, as long as these things are secure in their position, attacks against them are ignored as being silly.

Third, when an institution/belief/etc is being seriously threatened, then people react. And react precisely as one would expect people to react to a personal attack: impulsively, irrationally (as in, without pausing to consider pros and cons), and with force out of proportion. For example, see the inquisition — which only went all-out at a time when the Church was really being threatened.

Fourth, read the replies to your blog.

Well I completely agree with @Alejandro Java as a language has to grow and grow fast with newer languages such as Scala. But I think still that teams would prefer Java for a large enterprise projects rather than PHP or Ruby.

No language ever dies. Saying that Java is dead is like saying C died. Newer languages are replacing Java but it doesn’t mean that it is a “semicolon” for Java. What about Pascal ???

And JVM is surely not dead and will not die soon unless until someone would develop a platform good enough to replace JVM everywhere.

C# is not a successor to Java, it’s not cross platform, there is no official release of the .NET framework for Linux. Mono sucks, but it kind of runs c# in Linux, but until MS releases a Linux client then C# will never take Java.

In addition to mint and others http://www.shoeboxed.com uses Java too. I have not used scala, but I have used ruby on rails in high through put projects and scalability is a pain. The lack of multi-threading is a killer. IMHO the beauty of java has really been that it has been enterprise ready and what worked in 1.x worked in 1.(x+1). That is not the case at least with rails. Therefore at least in the Rails/Java debate, user rails for vc demos, but that is it.

I’m using Java 2D and Swing to write quite a good looking and responsive music notation and playback program

I think the “enterprise ready” comment is interesting. Have you worked in a large enterprise? What was the worst infrastructure they used? The reality is that enterprises put up with outdated junk as infrastructure that is orders of magnitude worse than anything anyone blogs about. Because they have large investments in that junk.

If the job of tooling is to allow developers to efficient create applications, then Ruby has better tooling that Java. Just write the same application with Rails and SpringMVC and see how long each one takes.

Deploy an application with Java on an enterprise Java application server, and do the same with PHP.

Jim,

What is your beef with Mono, is there something you tried to do with it and failed? What are your concerns?

I don’t a great deal of personal experience, but it is part of Gnome with two application written with it, and seems well supported and reasonably robust.

I tried writing a small application with it, and found C# a mostly improved Java.

I think you managed to nail it. The question is not “Is Java dead?” but “Is Java dead for me?”. For some that might be true. For others it might be the total opposite and Java is more prominent in their workspace than ever.

When Java started, it was REALLY ambitious. From enterprise servers to credit cards, everything should run Java. That era is over, and a sober attitude has settled, which I think is good.

The question could also be formulated; “Is the Java hype dead?”, and to that I think most people would answer “Yes”. It is no longer the sexiest thing you can work with, and it is these fashion programmers that are very vocal on their proclamation that Java is dead or that the language needs an overhaul.

I personally don’t think Java needs much overhaul (some, yes), but instead the many “wisdoms” and “patterns” that we are coding with are just plain wrong and greatly contributes to the “noise”, and this *could* be fixed within Java without the massive investment of a new programming language.

The investment side of Java is greatly under-estimated by the fashion programmers. The amount and quality of tooling, libraries and most importantly knowledge among the developers are immense, something that will take decade to replace. That alone should raise the question “Why?” and one should (as was the case of C++ to Java transition) look for evolutionary small steps and not massive replacement. Re-writes (so to speak) don’t pay off unless the benefits are 1-2 magnitudes greater, which none of the contenders can show at the moment (except for prototypes and smaller projects where maintenance crew equal one guy’s head).

“Java is dead” gives me 171 000 hits, not the millions you talk about. I think the mistake you did was to search for the term without the quotes around them. That just identifies pages with the words “java”, “is” and “dead” on the same page.

@stephan: When I argue for C# as the successor to Java, it’s obviously in a language capacity. Nothing changed in 2009 but C# is a superset of Java and fixed a lot of issues.

The fact that we do not have C# on the JVM is down to politics and religion, I mean you have every other arcane little language there. And over on the other side, Mono runs Java very well.

I am not saying C# is going to have to take Java’s place on the JVM, I’m saying that you’re leaving out an important piece of the puzzle in your list. Fan for instance (which happens to be my favorite) is heavily influenced by C#.

I can’t hear it any more – telling about “Java is dead”!

I did pretty intensive analysis about what should be my future in development and decided to go with Java.

Language features or verbosity is just one point. It also must be readable, easily learnable by the people coming after us, there must be libraries, good IDE, GUI designers etc.

I do quite each month read about a new programming language – today it was “Surinx” – see http://blog.headius.com/2009/08/introducing-surinx.html

I read a lot about Scala including a lot of things still not supported in IDEs and so on. I think, people try to reinvent the wheel again and again even without understanding why several older things have been implemented as they have been implemented.

I go totally conform with all the developers designing the Java language who think very well before implementing something new. It should not happen that such an important thing as the Java language become bloated and unstable. – Nobody thinks about C(++) introducing significant language changes.

I do software development since the early eightees and I have seen a lot coming along. Java is not dead, it’s just some “hypies” talking about it’s dead – nothing more! In my opinion talking about Java dying is just a hype.

I think it’s tough to compare these languages because they inherently attract somewhat different cultural groups.

Ruby/Groovy/JRuby and Scala are all pulling in the young, trendy folks who don’t have years of experience invested in Java on their resume.

C#, while having a solid crowd of hardcore windows platform devs is also heavily crowded with older VB programmers, who again, have a lot of time invested in “where they are” (read: Microsoft-teet-suckling). Unfortunately, that means a huge bunch of these peeps may be getting a lot of work done, but they can do it without really understanding OO. Look at the failure of webforms compared to the MVC model that’s come out somewhat recently.

Java on the other hand is still staffed to the gills with people who know what they’re doing because most of the frameworks are so wordy and over-thought that you have to know something about OO and patterns. I would argue that your Java guys tend to be a bit mightier in the CS background. So while the language is about cemented in stone and not very sexy in that regard, the community is still bringing plenty of good stuff.

I personally like the polyglot approach. Some languages like Ruby (with Rails specifically) are essentially domain specific and as such, you’re going to get a lot more bang right out of the gate. So why not use IronRuby or JRuby (or hell, just Ruby) for your web stuff, but use JavaFx or WCF+C# for your rich client.

Meanwhile, with web service contracts and other hooks, you can still share certain core functionality. Hell, you might even have some of that implemented in C for speed’s sake.

Use the right tool for the job instead of suffering vs a nail with yr tape measure. The polyglot approach also let’s you keep yr old timers rocking what they know and yr young turks happy with their shiny toys.

Just because there are better languages available doesn’t mean that they are going to replace java.

Look at history – COBOL and VB were hugely popular despite the existence of many better languages.

Corporations aren’t simply going to switch over their applications to a new language or platform just because it is better or even more popular; there is a huge investment in these legacy systems and it is almost always be cheaper and less risky to maintain them than to replace them with something written from scratch. It’s not just the initial investment in creating the software; it’s the infrastructure, support organizations, and procedures processes that exist around the software. These things do not just go away overnight. There are many companies out there running software on JDK 1.4 or earlier, despite the relative ease in switching to a newer JDK vs. switching to a whole new language or platform.

I agree with Casper Bang.

I just dropped a massive Windows application onto a Linux/mono2.4 and it worked great. After doing a little testing I did change a couple lines of code. But the process was as smooth, if not smoother, than the equivalent process for a java application.

Also, I’d like to clarify that java is a valid option. It isn’t going away any time soon. I see no reason why both stacks can’t co-exist. When you combine mono,SWIG, and IKVM, you have a great framework for using java,C#, and C++.

The TIOBE index shows Java is still on top at 19.4%

http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html

So far, the music notation app I’m developing, using Java 2D and Swing, works flawlessly on Mac OS X and Windows XP with no platform-specific tweeks. Music playback is perfect on my ancient 867 MHz Mac!

Interesting article. I don’t really agree with a whole lot of it though aside from your conclussion that Java is not dead.

Google around for most popular / most widely used programming language and all of the sites ive seen list java as #1. Here is one of them:
http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html

Far from dead. More along the lines of – ruling like a king.

One big thing I massively disagree with is that it was a huge mistake that java objects initial value is null. I think it is absolutely awesome. Exceptions are your friend. If something blows up – stack trace tells you exactly where the problem was. Fix it – good to go. Without exceptions you could spend hours to days trying to track down logic bugs.

Lets imagine that you wrote some complex math code and somewhere in there you forgot to initalize it (if you are using a good IDE it would flag it for you). You run the program – NPE. Stack trace tells you exactly where the problem is – fixed in less than a minute.

Now in a world Java world where all object had some non-null initial value: You run your code and oh… guess what? You didn’t get the output you expected. Or even worse you shipped it and then it causes major problems for the customer. Now you have to spend hours or possibly days trying to track down that damn logic bug.

No my friend, Java initializing objects to null is quite amazing and was a very intelligent design choice.

Now that Oracle is buying it, I think that Java’s replacement will be PL/SQL (don’t laugh just yet). Someone said “enterprises don’t care and will just buy the upgrade” and to an extent they are correct; at least some percentage will buy the upgrade to what ever Oracle offers. There is a large market of not very developer-savvy enterprises that buy all their stuff already. If you look at _nearly_ _all_ their “Java” Fusion Middleware products, under the hood there’s a stored procedure waiting for you !!! (ok you can laugh or cry now).

As for IDEs frameworks and “rapid” web development; I’ve had great successes with Tapestry 5, Jetty and Eclipse. You can edit the pages and reload the browser with restart and/or recompilation (eclipse of course is providing the incremental recompilation on save).

In the meantime, I think we are moving to a more polyglot universe. If you think about it, the ideas behind SOA (buy off Oracle) or REST (hooray!) are about enabling this sort of move by utilising standardised network transport layer protocols like HTTP and easily understood formats like JSON, HTML, XML etc. I think if you look at the “device enabled world” this becomes more apparent. I was just working on a project at a client that has a JBoss-enabled middle tier, with a quite large number of Windows CE devices “in the field” (ten of thousands) and a selection of .Net back office applications, all working happily together via web services.

“reload the browser with restart and/or recompilation” WITHOUT i meant WITHOUT of course.

As an academic experiment Java is dead. In some respects it worked (no pointers, platform independence, garbage collection) and in some it didn’t (verbosity, checked exceptions). Some initial problems were fixed, at least partially (concurrency, generics). But the outcome is basically worked out and the research is over, with the exception of adding closures, which will be resolved soon.

As a production language Java is not dead. The legacy code issue is enormous, and it’s not just big enterprises. I write programs in Java to use in my “real” job (biology), have been doing so since 1998, and I would probably need about 2 years to learn a new language and packages and rewrite everything. I don’t have 2 years for that, and it wouldn’t make sense in a cost/benefit sense anyway. I doubt I’m alone in this situation.

I think, in addition, Java will have a big role for open source systems. You need access to a huge public library for open source, and you need a language that’s very widely known. It’s tough to do open source when 97% of programmers can’t write in your language. Being propriety (C#) isn’t good either, both practically and philosophically.

I think Java will be fine in production until massive multiprocessing becomes routine. I recently wrote a system that can convert something that generates a series of items but can’t do next() conveniently into an iterator via multiprocessing. And it works great – except that the entire exceptions system is based around the assumption of a von Neumann system where if you throw an exception it gets caught by the calling code. When the calling code is in another thread it doesn’t work and the exception goes (inappropriately) to the top. Combined with the lack of list processing, I don’t think Java can work in a massively multiprocessed system. So in a decade or so Java will become defunct for new work although the legacy issue will remain for quite a while.

No arguments with the article (I dont know enough about the Java space anymore to comment), but your google fu is not correct:

—-
Searching for java is dead with Google, one gets

Results 1 – 10 of about 8,620,000 for java is dead.

Dead indeed. Or at least lots of people think it is or will die 2009.

—-

Sorry, but that text:

java is dead

finds any documents with the words “java”, “is”, or “dead” in it, or any combination. I get about 15m results. No surprise – is and dead are VERY VERY common words.

What you most likely want is:

“java is dead”

or maybe:

+java +is +dead

ie, the WHOLE PHRASE, with quotes. I get about 170,000 results from the first, and 1.5m from the second (google UK). A lot, yes, but not 15m.

??

[...] Is Java dead?Nicht vom Titel täuschen lassen! Dabei werden die Aspekte sehr sachlich betrachtet und durchleuchtet. [...]

I cannot wait for java being dead. Then I can enjoy its successor. But there’s no successor now!(Don’t tell me c#, I think Scala for .NET will be better, .NET is better but not c#, but it not goes far enough)
I will Erlang could be, because erlang vm is awesome. But the erlang language and Ericsson suck.
The problem is still concurrency. The successor must solve it and it will born for it! This is why I`m waiting for the successor.

think for it

but what don’t you talk about Php ?
I’m a java programmer, I go to zend framework and I find it very good to run web2.0 projects

I think php is the default choice for startups, and has most chances to become the next star : all the public excited projects are in php : CMS (drupal), ecommerce, multiblog…
it seems for me all the best opensource is in php.
too bad “old” developpers don’t think about it :)

No compelling server apps in Java – Hmm isn’t Adwords – the biggest money spinner on the planet done in Java – isn’t gmail done in Java? ( Java – to javascript via GWT ).

Isn’t tomcat and jetty java?

I think your quote about enterprise stuff being boring sum’s your view up – your not interested in shipping useful, valuable software, with perhaps more than a 5 minute life span – your interesting in playing with the technology for the technologies sake rather than what you can achieve with it – if you were more concerned with the end result then you may be less dismissive of Java.

When I see a resume with all the technologies under the Sun on there I see that as a danger sign of somebody who is more interested in playing than getting the job done.