Sun should manage the jcp process as they have been and allow anyone to contribute if they first commit a valid, useful jcp spec. this way sun can architect the language and many others can build it. it would blow .NET out of the water and it would save them time and money.

The real problem with Java client apps is the lack of widespread native compilers. Almost all of the attempted compilers projects were stymied by Sun's licensing strategy. If Sun went open source with Java, this would help these efforts greatly and revive developer interest. There are droves of developers that would love to help out with an open source Java.

C# is a very real threat to Java, as a lot of shops are consolidating their required skill sets to lower costs. If a language is suitable for both server and client side development, it becomes a more desireable alternative. Thus, the client issue is not going away and may have a direct impact on server market share. The vision of widespread OS's and hardware with built in JVMs has not become a reality either.

Sun really should consider the examples of Open Source projects that have centralized repositories managed by a group consensus. The examples I've seen, such as OpenBSD, have none of the "chaos" that some people assume must ensue with an Open Source effort. It all depends on the model you adopt for managing the product. Please Sun, imagine the vision of a centrally managed Open Source Java that has the ability to be run in a JVM or compiled natively on any OS in existance and encompasses the mindshare of thousands of developers around the world.

Sorry folks, not Los Angeles, but Latin America. So, if you think about it, take USA out and we are talking about 3 continents.

I feel OK using both C# and Java. I serve MS advocates and also IBM/Sun/Anti-MS advocates.

After all, I do not work for them, I work for my interests and those of my clients.

I'm writing some fairly large client applications in Java. The slow load time does not even come close to outweighing the advatanges of using Java, in particular knowing that the application will continue to run regardless of where the industry moves. If Linux starts to take over the desktop by 2010 (which I doubt) then I'm okay. If Microsoft continues to dominate everything then I'm still okay. If there's another surprise and some other OS springs to life by 2010 then it is almost certainly going to have a JVM written for it because of so many others having that requirement. The industry is obliged to continue the support for Java on various machines for obvious reasons. It is too much risk for me to write applications for which I intend to have a really long shelf-life, with .NET. Microsoft has a history of changing their language slightly over time and making developers to work to keep up. Sun have done a very good job of retaining backwards compatability to older versions of Java with their newer JVMs, and moreover you can include an older JVM with your application if required. Regarding slow load times of Java applications, I think this is just uninsightful because we are talking about fixed latencies which become negligable over the long-term as machines increase in speed. Loading a .html file in the late 1980s would have taken minutes because the browser would take ages to load, but now it is just an instant thing.

Is at server side. Client side Java is hard to compete with MS technologies due to the dominance of Windows. On my PC, Java Swing application is even getting slow when I upgrade from win98 to Win XP.

But MS should be aware of the competition of Linux as well. Look at what is happening in Europe and Asia, you will understand what I mean.

... sounds like a good plan

Steve - as David Geise indicates, IBM is indeed making efforts in this area. Please check out the *open source* Jikes Research Virtual Machine (Jikes RVM) at
http://oss.software.ibm.com/developerworks/oss/jikesrvm/?origin=jikes

This is a pretty impressive effort in this direction but would really benefit from more folks downloading and trying it out and otherwise getting involved. IBM's backing of efforts like this will be proportional to the interest demonstrated by the rest of the world.

While Microsoft is out spending SUN, has anybody spoken to IBM, to see if they were interested in putting some muscle behing open source JAVA. IBM has the potential of overpowering Microsoft resources. Remember although IBM and Sun are competitors, Microsoft is a much bigger competitor and, I am sure IBM has not forgotten what Microsoft did to them with the 1st OS/2 release they wrote for IBM.

Take a look at the Jikes Reseach Virtual Machine (http://oss.software.ibm.com/developerworks/oss/jikesrvm/?origin=jikes). Its written in Java and is self-hosted (doesn't require another virtual machine to run it).

Referencing the related post by Roger Dawson at 6:29 PM yesterday... I read somewhere very recently that the main focus of 1.5 is to vastly improve JVM startup time. The author -- can't recall who -- spent some time at Sun and was apparently impressed with the demonstrated improvement. So... may be there is hope. [I wonder if there is a way to 'freeze-dry' a JVM at some point during it's initialization, and save it for later re-load. Probably not a new idea, and likely more easily said than done... but why not think out loud anyway.]

Pat, it seems to me that you don't have too much of an idea on what any company needs to outspend another. When you speak of the Java language itself, there's no need to spend a lot. A single university department is able to create a 10 times more advanced language than Java or C# will ever be, within a few months.

What needs money is the infrastructure around the language and marketing of course. And there you have to compare the whole of IBM, Sun, Oracle, BEA and lots of others against Microsoft.

IBM and Sun have one tremendous advantage over MS. They are not dependent on software license sales. IBM will keep pushing free open source software into the market and earn money on consulting and hardware. Microsoft will have to struggle a lot to find a means against this business model. It's just not theirs.

I'm, with reluctance, switching to ASP.NET, C#, ASP.VB, MS SQL Server and Visual Studio. I will even take a look a the new MS Server...

It just seems to me that Sun can not compete with Microsoft. As Mr Gosling pointed out, MS is outspending Sun 10 to 1 in the language area. It seems like Java going to Open Source is the only solution but I just can't seem to see Scott turning it over to them -- hope I'm wrong here...

C#, in my opinion, is really a pathetic language but it is compact and easily learned and apps written in it can be readily implemented. I also tried ASP.VB and found it, especially when using Visual Studio, to be easily learned...

Perhaps if Sun tried to compete against Microsoft in the development lab rather than in the court room, things would have turned out better. I say this as a 'badly burned' Sun stock holder.

Good luck to us all,
Pat

I agree with you. JVM startup time is too long. But it's not enough for me to justify switching to .NET. I know Sun's team is working hard to optimize performance of Swing applications, even if they are slow to do it. I think if Java had been Open Source, this issue would be resolved more easily by the community itself.

I had said that Java will one day become open source. That was 3 years ago. I even wrote an article for Enterprise Linux Magazine about Java on Linux is an excellent option for open-source initiatives. Back then, I was still working for Sun. I guess the management finally listened to my ideas!

I had to switch to C# reluctantly because the load time of Java desktop GUI applications took in the range of 25 seconds or greater. Granted C# load time is slower than a C++ or Visual Basic but is a magnitude faster than the load of Java. Once loaded there really is no difference in the reponse time of Java vs C#. I had to switch because the Java load time was excessive. I again repeat that once loaded the execution time was just as fast as C# or even C++. I could never understand why Sun did not address this disparity. It makes me feel sad when I think about it. I really did not want to use a Microsoft invented language.
I feel sometimes like I am the only one with this complaint. Surely with millions of developers using Java someone else must have the same problem of excessive load time for Java desktop GUI apps.
Regards
Roger