a Very tough question: what's Java's future

[Will Lee]

I can't help wondering, can I still code in Java in 10 (or 5) years?

[Paul Clapham]

Yes.

[Pat Farrell]

no. At least not me. All languages have a life cycle. They start, get popular, and then lose popularity. Some stay in use a very long time after their popularity peak, others are lost and forgotten.

Snobol and Smalltalk come to mind as cool languages that faded hard. Fortran and Cobol are still in use after 50 years.

But I think Java's age is showing, and its not well suited for use on AMD's new 16 core CPU.

[Henry Wong]

But I think Java's age is showing, and its not well suited for use on AMD's new 16 core CPU.

What makes this processor bad for Java?

Henry

[Pat Farrell]

Henry Wong wrote:What makes this processor bad for Java?

Its not this particular processor that is bad for Java. Its Java is bad for that processor. Or, rather its that while Java has Threads, (see Henry's excellent book), its too hard.

When we had single processors, it was OK to use threads manually to allow for responsive GUI, etc.

But its just too hard (IMHO, YMMV) to manually decompose your application, fix the race conditions, etc. When commodity computers have lots of cores, mainstream developers are going to need the language/compiler/runtime system to do it for them. Its already clear with the common quad processors, going to 16 is just going to make it worse.

I don't know what the future answer will be, as writing correct parallel programs is really, really hard. But I don't see Java as the tool to do that. Perhaps Scala, perhaps something completely new.

[Henry Wong]

But its just too hard (IMHO, YMMV) to manually decompose your application, fix the race conditions, etc. When commodity computers have lots of cores, mainstream developers are going to need the language/compiler/runtime system to do it for them. Its already clear with the common quad processors, going to 16 is just going to make it worse.

I will try to answer this in two ways...

First, I'll agree that threading is a skill -- and some developers just can't seem to get the hang of it. This is why application servers are popular -- it is the application server that deals with the threads. These developers deal with their component -- and the application server scales it for all the users that access the component, and for the different components that access each other.

Basically, let the application server deal with threads....

Second, for developers that can deal with threads, Java has proven itself in this regard. I have worked on a Java application that ran on a machine with 384 processor cores -- and we kept 300 of those cores busy processing data. We probably would have kept more busy, if we had bigger thread pools...

In this case, it was the profiler that made it easier. We had the tools that helped with finding race conditions, and locate deadlocks, etc. This profiler also did a bunch of other cool stuff....

I don't know what the future answer will be, as writing correct parallel programs is really, really hard. But I don't see Java as the tool to do that. Perhaps Scala, perhaps something completely new.

I think the future is a combination of the two points that I mentioned. Application servers or framework that deal with threads. And tools that make it easy to develop and debug parallel applications. Changing the language may be possible too -- but that strikes me too much like throwing the baby out with the bath water.

Henry

PS... Thanks for the plug ...

[Rob Spoor]

Pat Farrell wrote:

Henry Wong wrote:What makes this processor bad for Java?

Its not this particular processor that is bad for Java. Its Java is bad for that processor. Or, rather its that while Java has Threads, (see Henry's excellent book), its too hard.

When we had single processors, it was OK to use threads manually to allow for responsive GUI, etc.

But its just too hard (IMHO, YMMV) to manually decompose your application, fix the race conditions, etc. When commodity computers have lots of cores, mainstream developers are going to need the language/compiler/runtime system to do it for them. Its already clear with the common quad processors, going to 16 is just going to make it worse.

I don't know what the future answer will be, as writing correct parallel programs is really, really hard. But I don't see Java as the tool to do that. Perhaps Scala, perhaps something completely new.

If you see Java die out because of this, then say goodbye to C, C++, C#, etc. You can write correct parallel programs in almost any language, but for most existing languages you still have to do the hard work yourself. There may come libraries to make life easier for you but in the end it's still mostly up to the programmer.

By the way, you say that writing parallel programs is hard in Java, but with java.util.concurrent, java.util.concurrent.atomic, the executor framework and inclusion of SwingWorker it's never been easier writing parallel programs in Java.

[Pat Farrell]

Rob Prime wrote:By the way, you say that writing parallel programs is hard in Java, but with java.util.concurrent, java.util.concurrent.atomic, the executor framework and inclusion of SwingWorker it's never been easier writing parallel programs in Java.

I don't have any argument, its easier than ever. But its still too hard. Its hard in any language, and much harder in procedural languages (C, C++, ObjectiveC, Java, Cobol, etc.).

We have today $500 desktops with quad processors, that AMD has announced 16 processors should surprise no one. with 32 and 64 processors clearly within a few years.

IMHO, Java isn't going to cut it. YMMV.

[Bear Bibeault]

What do you see as the compelling alternative?

[Praful Thakare]

As Rob correctly mention, Application servers and frameworks (or technologies) like ejb is better choice over writing multi threaded java program.
y not use 100 stateless session bean pool than writing one class with 100 threads?

Also what does IMHO and YMMV stands for ?

[Pat Farrell]

Bear Bibeault wrote:What do you see as the compelling alternative?

Not sure who this is addressed to. But me, I don't see any compelling alternatives yet. But necessity is the mother of invention.

If you had asked five years ago, I'd say something more pure OO than Java (no int/float, etc.) but I think the OO fad has run its course. 30 years is a good run, and Smalltalk 80 was the first I worked on.

[Pat Farrell]

Praful Thakare wrote:Also what does IMHO and YMMV stands for ?

Google or urban dictionary, or hacker's dictionary will define it for you. Better than us just telling you.

[Pat Farrell]

Praful Thakare wrote:Application servers and frameworks (or technologies) like ejb is better

All of those things have nothing to do with Java. That Java needs them indicates that the language is not a good basis set for the problems.

I'm willing to bet that no one will be writing EJBs in ten year. Period. Other than folks supporting legacy implementations. I don't think EJB ever really worked.

[Paul Clapham]

Pat Farrell wrote:Other than folks supporting legacy implementations.

And this is why I answered "yes". Millions of lines of Java code have been written and even if people stopped writing Java tomorrow -- which they won't -- there would still be the need to support that code for years to come.

[Pat Farrell]

Paul Clapham wrote:And this is why I answered "yes". Millions of lines of Java code have been written and even if people stopped writing Java tomorrow -- which they won't -- there would still be the need to support that code for years to come.

By that definition, I wonder if any computer languages have actually died?

I assume that the binary machine codes of the Eniac died when the last Eniac was unplugged. But I know that there is a fair amount of PDP-10 Macro assembly language still being executed and even occasionally patched, and the last PDP-10 hardware was made in 1984. Some of the design concepts of the PDP-11 RSX operating system were included in VMS for the VAX, and NT was built by the same people, with some of the same techiniques as VMS, so it had PDP-11 RSX concepts in it. We are, of course, using a direct descendant of NT today in XP and Vista.

I've seen this question on every forum/bbs/mailing list that I'm on that touches Java. I find the whole topic silly. I expect its popular because lots of folks consider themselves as "Java developers" and worry that if Java dies, they will lose their jobs. First, as @paul says, there are millions of lines of code that will need support. Second, you should consider yourself a "professional developer" rather than a "java developer". You may just happen to use Java today, but if you are smart enough to be any good, you are smart enough to learn python, perl, pf php.

Languages change, grow, and die. Get used to it.

[Anila Mathew]

Its true.

But I also think that its natural to be uncomfortable considering you have worked on this language for so long, and you kind of feel its the best so far.

For me, Java was just not a programming language- It was a platform where so many other technologies / frameworks were formed upon.

I know, it might be silly to say that I would not work on any other programming language in future, but I hope whatever comes new , will have the flexibilty and ease of learning like Java.

[Pat Farrell]

Anila Mathew wrote:any other programming language in future... will have the flexibilty and ease of learning like Java.

All languages are flexible. Its been proven that you can write any system in any language. Just some languages are easier than others.

Folks in this thread have talked about a lot of things, as solutions, that are nothing close to easy to learn. EJB, Servlet servers, and frameworks are much harder to learn than the Java language itself. Java is nearly trivial, even with all the cruft that has been added over the past decade. But you can't convince me that all the EJB, DAO, and other stuff that take Java from a simple set top box implementation language to a language for enterprise implementation system, are easy to learn.

[Christophe Verré]

Google or urban dictionary, or hacker's dictionary will define it for you. Better than us just telling you.

Please use real words. I don't understand most of these abbreviations either.

[chris webster]

Anila Mathew wrote:
I hope whatever comes new , will have the flexibilty and ease of learning like Java.

Or you could hope that whatever comes along will be even more flexible and easier to learn than Java!

Grails (mmm....lovely Grails...), Python, some other all-purpose language, who knows - but it'll be fun finding out.

[Rizvan Asgarov]

Hi all...
Firstly Java is not behind than others... Because of platform that is independent is a good reason when looking at others...
If Oracle gets it indeed so i think they will develop the joint application of these.
But if they are different thoughts... So doomsday will not allow these!

[Arun Kumar]

If a language has to stay it shoudl have some qualities which are better than others

now for what application will you use JAVA in future ,if its just web development PHP might be the better solution

so I think the best way to find out whether java will be around for some more time will be to define what all areas java is still better than other languages

[Paul Sturrock]

Arun Kumar wrote:
If a language has to stay it shoudl have some qualities which are better than others

I'd disagree. It needs to be usable and feature rich. But a big community or corporate backing will have a far greater effect than a language's qualities (otherwise this forum could quite possibly be SmallTalkRanch.com).

VB for example is not a very pretty language, but there are still loads of VB programs and programmers out there.

[chris webster]

Arun Kumar wrote:
If a language has to stay it shoudl have some qualities which are better than others...

The problem is in defining what you mean by "better". Smalltalk is a "better" language than Java in OO terms, C is a "better" language than Java in terms of efficiency, Python is a "better" language than Java in terms of being easy to learn, PHP is a "better" language than Java for quick (and dirty) implementation of web applications, and so on.

People's ideas of what is "better" depend on lots of factors - including the availability of skills and technical support for the expected lifetime of the system, as Paul says - and may change as technology develops e.g. people are far less concerned about squeezing the optimum use of memory or storage space out of many systems these days, as it is often cheaper just to buy an extra terabyte of storage or an extra GB of RAM than spend an extra 3 months developing a slightly more efficient application.

Java will last as long as there are enough people who feel it is "better" for what they want to do with it. Then it will be replaced by something, well, "better"!

[Henry Wong]

For some reason, I am reminded of the quote from Bjarne Stroustrup...

There are only two kinds of languages: the ones people complain about and the ones nobody uses.

This was quoted during the period where C++ was poised to become the premier object oriented language. And ironically, it was probably a response, in frustration, to all the people who are complaining, to just "get over it".

The fact that people are complaining, is a sign that Java is popular. Hopefully, Java will keep those complainers, and get more.... Not by not listening, but by improving based on those complaints. Getting more fans. And in the process, getting more complainers...

Henry

[Will Lee]

chris webster wrote:
Java will last as long as there are enough people who feel it is "better" for what they want to do with it. Then it will be replaced by something, well, "better"!

That is what I am concerning. It's all engineer talking here. How about business talk? After Oracle acquired SUN (and Java, even though people will argue that SUN doesn't actually own JAVA), are they going to change the track that Java is currently on? Many of Oracle's rivals use Java in their products (I heard SAP is one of them). How about follow Microsoft's step: make your project (which is Java in this case) hostile to your competitor? Then the industry will drift away from this language. Perhaps nobody is going to complain about Java anymore at that time .

Is this going to happen?

Maybe it's time to join .NET?

[Pat Farrell]

Henry Wong wrote:This was quoted during the period where C++ was poised to become the premier object oriented language.

Great quote. And contrary to expectations, C++ never became the popular OO standard.

IMHO, that was because it was too complex, the OO stuff was just layered onto C, and it was not elegant. Too many weird constructs, const refs to const pointers, etc. Interesting, because the father of OO, Smalltalk was simple and clean. The syntax could be learned in one sitting, and nearly everything that you wrote would work.

One language that would trivially support the coming 16/32/64 core processors was Iverson's APL, because it was declarative, rather than procedural. The APL language, of course, was a complete failure. Kleenix code, write once, throw it away.

[Bear Bibeault]

Will Lee wrote:How about follow Microsoft's step: make your project (which is Java in this case) hostile to your competitor? Then the industry will drift away from this language.

Maybe it's time to join .NET?

These two quotes seem to make no sense together. "Oracle might make Java hostile, so let;s move to a platform that's already hostile".

[chris webster]

Will Lee wrote:
How about follow Microsoft's step: make your project (which is Java in this case) hostile to your competitor? Then the industry will drift away from this language. Perhaps nobody is going to complain about Java anymore at that time .

Right now there are billions of lines of Java code out there, millions of programs, thousands of companies using Java for all kinds of applications from company-wide JEE systems to mobile phones, tens of thousands of developers writing Java applications like these, thousands working on all kinds of valuable open-source products based on Java, and probably thousands more who will be moving onto JVM-based languages like Groovy in the near future.

Now imagine you already use Java for many of your own key products and internal systems, and you just paid $7 billion to buy the company - and experts - that created, owned and maintained Java.

Do you want to be:

a) The top Java authority in the world, the "go-to guy" for All Things Java, and the key player in determining the future direction of Java?
or:
b) The Company That Killed Java without having anything better to put in its place?

It makes no business sense for Oracle - or anyone else apart from Microsoft - to kill Java any time soon, unless they can persuade the Java world that they have something better to replace it with. And .NET is hardly an alternative if you're worried about depending on what a big software company plans to do with its proprietary software platform.

[Paul Clapham]

I think it's worth mentioning that none of the Oracle-Buys-Sun-What-Now posts which I have read, and which were written by people who understand the industry, even mention Java. They are all talking about databases, and servers, and virtualization, and things which people actually buy and pay lots of money for.

So I think that imagining evil villains who want to kill Java, or caped super-heroes who want to save it, is just nonsense. In the board rooms it's hardly even going to make it onto any agenda.

[Pat Farrell]

Paul Clapham wrote:that none ... even mention Java. They are all talking about databases, and servers, and virtualization,

I think you are right in spirit, but I have seen occasional mentions of Java.

Java does not compete with any of Oracle's lines. So there is no reason for Evil Oracle to kill Java and remove all references to it in history.

There is much more room for Evil Oracle to kill MySql.

[Brian Legg]

Will Lee wrote:I can't help wondering, can I still code in Java in 10 (or 5) years?

Yes. As others have said there is a ton of Java code out there that will not be going anywhere anytime soon, not to mention that new code is still being created and there is no reason for now for that to stop either. I think that over time programming languages have become easier and easier to use/learn and if Java is replaced then I am looking forward to how slick the new language will be.

My only complaint with Java.... I can't find any work I've been doing MS crap for so long now and I hate it... I miss doing Java professionally.

[salvin francis]

Brian Legg wrote:
My only complaint with Java.... I can't find any work I've been doing MS crap for so long now and I hate it... I miss doing Java professionally.

wow you are missing it already,

Dont worry, with MS coming up with new versions of its own languages every year, you will soon be forced to switched to java

MS is extremely good at developing application that "seem" to deliver a lot, rather has a lot of cosmetic sugery done to make it look good,
but in the end, it suffers in code maintenance.

afaik





Meanings:

MS: Microsoft
afaik: As Fas As I Know

[salvin francis]

Will Lee wrote:I can't help wondering, can I still code in Java in 10 (or 5) years?

The answer is same as the Question

Will i still enjoy chicken after 10 years?



Answer:

Yes, but you never know what the future has in hand.........

[Paul Sturrock]

Dont worry, with MS coming up with new versions of its own languages every year, you will soon be forced to switched to java

Doesn't Sun release new version of Java every year?

[Prabhat Gupta]

I little bit know palmistry ...i would see JAVA hands then only augur about its fate

[Henry Wong]

I have to admit, everytime I see topics like this, that I think "why would anyone want to focus on any one language"? Maybe it is because I don't work for a "pure Java shop". My previous company was a "pure Java shop", but like any company, they have a small number of C++, C#. Javascript projects from time to time -- and I volunteered for all of them.

Now, don't get me wrong, I probably use Java more than any other language. It is definitely the language that I am most comfortable using. And if I must choose to work in only one language, Java would be my choice... But... my question is... Why would anyone go out of their way to restrict themselves? It may be comfortable. And hence, it may be enjoyable. But heck, isn't trying out and learning new stuff enjoyable too?

Henry

[Cameron Wallace McKenzie]

I know some COBOL programmers who are still active. Regardless of what happens in the industry, Java will be around for a long time.

-Cameron McKenzie

[Pat Farrell]

Henry Wong wrote:. Why would anyone go out of their way to restrict themselves? It may be comfortable. And hence, it may be enjoyable. But heck, isn't trying out and learning new stuff enjoyable too?

Agree. Its good to try other languages and tools, just to expand your brain. I consider myself a professional developer/designer, not a Java developer/designer. I have used Java for at least 90% of the code I've written in the past decade.

To me, a much more interesting question is: "what are the Snobol and Smalltalk of the next decade?"

As both of these were ground breaking languges. Neither had the impact of a Fortran, COBOL, or C.

[salvin francis]

Paul Sturrock wrote:

Dont worry, with MS coming up with new versions of its own languages every year, you will soon be forced to switched to java

Doesn't Sun release new version of Java every year?

let me rephrase:

Dont worry, with MS coming up with new versions of its own languages every year That are Incompatible with the previous, you will soon be forced to switched to java

[Clayton Burke]

This is an interesting topic and set of responses. I'm new to Java, but experienced in several other languages (c/c++, perl, python, fortran, php, ...etc). I've avoided Java (probably subconsciously) all this time because of my initial disdain for the typical jvm implementations in the 90s. They really tuned me off for various reasons and left a lasting bad impression of Java (which wasn't really deserved).

I started using Java about 2 years ago by way of Google Web Toolkit. I had been writing a lot of Javascript based web-apps and I was getting frustrated with cross-browser compatibility issues and GWT seemed like just the ticket. Even though I'd be writing in Java, it was getting translated to Javascript - so I didn't have to worry about the end user's jvm. Well, over the course of the next 2 years, as I continued to develop GWT apps, my respect for Java grew to quite a high level. In fact, I even told my wife on one occasion that learning Java actually changed *me* in the process. This didn't really happen (or happen to a noticeable degree with the other languages I've worked in). Reflecting on this, I've come to believe I feel this way because of the myriad of very long thought out concepts implemented by Java or implemented in Java, many of which sent me on hours (sometimes weeks) of deep thought when I first 'discovered' them for myself.

So whether Java remains active for decades to come or fades out as new languages emerge - the new ideas it fosters and the impression it's made on the development community will surely outlast the language itself.

Thanks,

C