In my opinion, this is the beginning of the end of open-source freeware technologies.

http://www.theserverside.com/news/thread.tss?thread_id=40611

Java Succumbing to .NET in my Organization

I work for a federal government contractor outside Washington, DC, and I am part of a small group within my organization that produces small- to mid-scale applications, mostly websites hitting a database, for our clients. The management of my group, who are not technologists, has decided to embrace .NET wherever possible for our applications. They cite that Java applications take longer and run into more significant problems than their .NET counterparts.

After talking with some of the developers who are in the trenches, most of them are happy to move to .NET. This even includes some who, like myself, have been certified in and built a career on Java. Apparently, Java is just too hard and too intimidating, especially for beginners.

Java and .Net are comparably powerful technologies. As one who has experience with both, I will say that I prefer the Java experience (especially for developing business logic), so I find it discouraging that Java is losing ground in this way in our organization. I have thought about the reasons why and have come up with some reasons (and I acknowledge I am probably not the first to offer these explanations):

· Too many frameworks

If I want to design a website look-and-feel and page flow in .NET, I use ASP .NET. That’s it. If I do so in Java, I use JSP. Or I use JSP and JSF. Or Facelets and JSF. Or Echo2. Or Tapestry. Or Velocity. The number of choices can be intimidating. And that is just for web design. With persistence frameworks, .NET developers have ADO .NET. Java developers have entity beans and JDO as Sun standards or open-source options like Hibernate, Torque, Spring JDBC, and on and on.

Given all the choices for the web design, the middle-tier, and the back end, developers are overwhelmed by the number of permutations of possibilities. I could go JSP/JSF, Spring, and entity beans; or Tapestry, session beans, and Hibernate; or JSP/JSF, session beans, and entity beans; or ... You get the idea.

· Too many IDEs... all with weaknesses

Closely related to the issue of framework proliferation is the issue of IDE proliferation. Consider IntelliJ IDEA, my favorite IDE. It is an amazing piece of software, but it was really behind the curve on JSP development. And that is a Sun standard! IDEA is also completely unaware of JSF, another Sun standard that has been around quite a while and one whose goal was primarily to facilitate web development in IDEs. If IDEA has this much trouble with Sun standards, imagine how it does with open-source solutions like Spring and Hibernate. Sure there are workarounds and plugins and so forth, but those take time .NET developers don’t need to take.

Of course, if, for example, developers want to use JSF, then they can go with NetBeans, which is offered free from Sun. But then they forego the incredible refactoring capabilities and “smart coding” capabilities that make IDEA so great. For Java developers, it becomes a matter of deciding which IDE is the least incomplete for their needs. Developers in .NET have no such decisions to make.

· Too many infrastructure component possibilities

Developers in .NET have Windows Server and IIS. That’s it. All development organizations have professionals devoted to these technologies. Java developers have Tomcat or Apache, for example, for the web server. If they need to deploy to a full-fledged J2EE server, then there’s WebLogic, OC4J, JBoss, and on and on. Developers need infrastructure staff trained in the organization’s preferred server, and the staff might need to be retrained should this change. In my organization, there is no such staff dedicated to J2EE infrastructure as there is with Microsoft infrastructure.

It should be noted that deployment to IIS is a simple matter for developers—as simple as copying the project to the wwwroot folder. Java developers must build WAR files or EAR files in order to deploy across servers. Not terribly difficult thanks to Ant, but still more work than a simple copy.

What bothers me the most is that these problems negate two distinct advantages that the Java community has over .NET:

· Open-source energy

Microsoft has a lot of talented people working to improve .NET, but no single entity is capable of the creativity of many entities. There is also the possibility that Microsoft creativity may be somehow subject to the company’s corporate aims. Both of these factors can potentially limit the ability of developers to deliver the highest quality software in the least amount of time.

Because no one quite owns Java, much to the chagrin of Sun, there are a lot of great open-source tools out there like Spring, Hibernate, and Struts that result from the efforts of many brilliant, dedicated individuals. The Java community should be able to benefit from this universal effort, and it should be an advantage over .NET. Yet instead it has become a detriment.

· Manifestation of agile methodologies within tools

The Java community embraced agile methodologies well before .NET did (if it even has). Tools created to facilitate agile practices like JUnit, Ant, and CruiseControl, all of which are open-source incidentally, were developed in Java for use in Java projects. Moreover, many leading IDEs like IDEA embraced these tools as core components. The result was a generation of IDEs that served as a one-stop shop for coding, configuration management, automated unit testing, and automated builds.

The .NET community, on the other hand, lagged behind the Java community in all these areas except for configuration management. And even there, Visual Studio .NET only integrates with Visual SourceSafe, another Microsoft product, rather than the all the other choices like ClearCase, Subversion, etc. In the area of other tools related to agile methodologies, only after some time did the fledgling open-source community on the .NET side create equivalents—NUnit, NAnt, and CruiseControl .NET—to their Java counterparts. Agile methodologies transcend implementation technology, yet Microsoft seems to not be quite on board. Only the most powerful, and most expensive, edition of Visual Studio .NET 2005, has NUnit integrated. No such luck with NAnt.

As with open-source energy, the agility advantage has not proved to be a boon for Java. This is because the business community has not grasped the business value of agile techniques—particularly in my organization. Therefore, the advantage Java has in this regard is meaningless.

Again, the technologies are comparable, so the issue is development speed and ease. Ultimately, I think it comes down to the IDEs. If Java development tools allow for rapid development and help to ease a lot of the pain associated with configuration and such, then this would narrow if not close the gap with .NET. Visual Studio 2005 has a number of really compelling features that make development fast and easy and far more palatable to my management. It is no small challenge, however, for the Java realm to accomplish such a feat with the myriad of frameworks out there.

Perhaps the answer is IDE standardization. JSR-198 is a very weak start but a start nonetheless. All IDEs could be built with some minimum standard functionality—like support for all production standards from Sun like JSF and JDO and standard development functionality like drag-and-drop web design. Another standard capability, as Eclipse has pioneered, would be plugin integration and a standard interface for development with the plugin inside the IDE. It would be the responsibility of open-source framework developers to build their software so that it can be incorporated into the standard IDE according to the rules for plugin integration. In this way, Java developers could leverage the power of the open-source movement and agile methodologies while maintaining the simplicity enjoyed by .NET developers.

I would love to hear your thoughts on this topic. How can I convince my management not to abandon Java?

"...but no single entity is capable of the creativity of many entities..."

The whole thought process that more is better just gets lamer and lamer every time I hear it.

ChrisA wrote:

If Java succumbs to .NET it will be a true shame since Java is basically, almost open source.

.NET is more open than Java.

.NET (well, the specification) and C# have been submitted and ratified by the ECMA, an independent consortium/standards body. Java hasn't, and remains under Sun's complete control.

ChrisA wrote:

If Java succumbs to .NET it will be a true shame since Java is basically, almost open source.

Java is sadly not open source yet. Sun anncounced at their recent JavaOne conference that they are will be start taking Java down the open source road but it's difficult because it means untightening the legal framework that surrounds Java, so it could take years.

Meanwhile IBM, BEA, IONA, Software AG, Sonic et al are busy working on an alternative to J2EE here in the UK at IBM Hersley (near Winchester, Hampshire)

Java is loosing ground because of J2EE and how it sucks. JBOSS was about the only thing that made it good. ... and Visual Studio 2005 is easy to use and very productive ... and us developers love easy lives and we'll sacrifice principles for ease because we are tarts!

How would Java benefit from being open source?

ChrisA wrote:

Not controlled by any entity, hundreds of thousands of developers can help to improve java as a language, long term survival of the language should Sun go under as a company, formal support by Linux companies in their products.

But .NET isn't either.

Java will become more important on the Unix-based platforms while it begins to fade away on Windows because of WinFX and Visual Studio.


Regards,
Vincent

But you can add classes and tons of other stuff to .NET. You can build on top of it, make your own namespaces and have them do whatever you want. Does Java not have the same capabilities.

Can you imagine thousands of developers "adding functionality" to a language as big as Java or .NET? Who would control the mess and organize it all?

littleguru wrote:

......
Simple types and class types. Why is an int not a class why do I have to put it into an Integer to be a class. JAVA implemented to much, to fast. With going open this is going to happen even faster with less thought and less general concept.

I regard Java as a respectable opponent of .NET.

I remember at the beginning when .NET adopted "everything is class" notion and decided to make even int as a class, some critics pointed that will incur performance penalty due to the boxing and un-boxing. I haven't known the final answer of that discussion. Personally I feel int is so basic data type, why bother to make it as a class? Can some insiders here shed your great light on this issue?

ChrisA wrote:



W3bbo wrote:  ChrisA wrote: Not controlled by any entity, hundreds of thousands of developers can help to improve java as a language, long term survival of the language should Sun go under as a company, formal support by Linux companies in their products. But .NET isn't either.

Yes it is, its controlled by Microsoft.  Can I see the source code to the .NET Framework?  Can I modify it and redistribute it? As long as I pass the compatibility tests with Java I can do that with Java.

And there lies the problem, it's very difficult to pass the compatibility and complicance tests in J2EE, Sun has moved the goal posts and compatibility with each release which has p*ssed off the likes of IBM.

Also the reality of Java being open source is that thousands of independent developers won't be coding for it, thousands of developers in IBM, BEA, etc will be. Make no mistake other corps will control Java.

ChrisA wrote:



W3bbo wrote:  ChrisA wrote: Not controlled by any entity, hundreds of thousands of developers can help to improve java as a language, long term survival of the language should Sun go under as a company, formal support by Linux companies in their products. But .NET isn't either.

Yes it is, its controlled by Microsoft.  Can I see the source code to the .NET Framework?  Can I modify it and redistribute it? As long as I pass the compatibility tests with Java I can do that with Java.

This is actually what's wrong with much of Open Source.  The source is the documentation, the design and the specification.

A standard is a much better specification.  If there's a bug in Sun's implementation, in order to "pass compatibility tests" I've got to implement the bug.  Then we they fix the bug, I'm no longer compatible.  Granted, a specification can have flaws as well but the types of flaws are very different (you won't find an overflow or off by one error in a specification).

I'm not making the argument that Open Source is bad.  I've been involved with (both consuming and producing) Open Source for years.  But black and white arguments like yours just don't work.

Java was the worst of all worlds before... closed source, no standard and yet "designed by internet/committee/participating corporations".  If they move to Open Source it would be a step in the right direction, but for languages I'd much rather have them standardized.

ChrisA wrote:

When Java is Open Sourced, .NET will lose share.  Java is a much better language

I beg to differ, C# is way more flexible and polished (and that's just the language)

ChrisA wrote:

I write once run anywhere

For simple Java apps, yes. But for more complex apps then you run into difficulties, especially with GUI widget libraries.

ChrisA wrote:

and in my opinion, Java is a true standard.

Java isn't standardised.

ChrisA wrote:

Can I take an application compiled by the .NET Framework and run it on Mono?

Yes

ChrisA wrote:

This is where Java excels

Show me a "pure" (no platform-dependent assemblies) v1.1 .NET application that doesn't run under Mono

ChrisA can drive every post in which he appears to be a mess and flame-warring place.

Shame on you guys who feed the troll. Don't ruin any post anymore.

Note: I am writing this as a big fan of C# and the .NET framework, and one who is not particularly fond of the Java language. I am not very familiar with various Java technologies except as an overview, so correct me if I make any mistakes or assumptions.

Original Article wrote:

Too many frameworks

Here I agree and disagree. Most Java web-programmers I know are always bemoaning the proliferation of website frameworks, none getting enough attention to fully mature and cover all options. ASP.NET covers the low-level (caching, HTTP piplelines and so forth) up to the presentation layer itself (WebForms, etc). Even there, though, there are many lacking features. It's a sign of .NET's maturity that alternate packages like Atlas or Ajax.NET are building on top of ASP.NET to add more functionality. The many Java frameworks, from what I gather, usually invent the wheel a lot more than they need to.

On the ADO.NET front, however, I don't see how the ADO.NET libraries can compete with a serious database layer. ADO.NET just provides the basics, lacking a lot of important features. Any mid- to large-scale project uses the Enterprise Library/DAAB at the very least for data access, sometimes going for more elaborate storage or O/RM solutions. The very examples he gives in the Java world - Hibernate and Spring - are being ported and gaining momentum in the .NET world as NHibernate and Spring.NET.

On the whole, I agree that too many options can be bad, but when you have a strong base, it's good to have options to tweak, customize or enhance.

Original Article wrote:

· Too many IDEs... all with weaknesses

The situation is similar here. Give me an excellent IDE for the base, and give a good extensibility model. Visual Studio is wonderful, but it has many holes. People like JetBrains, makers of IntelliJ IDEA, plug into it with Resharper and make it even greater.

I have to say I've been annoyed with Resharper 2.0 on VS2005 though. Too goddamn slow so as to be unusable.

Original Article wrote:

Because no one quite owns Java, much to the chagrin of Sun, there are a lot of great open-source tools out there like Spring, Hibernate, and Struts that result from the efforts of many brilliant, dedicated individuals. The Java community should be able to benefit from this universal effort, and it should be an advantage over .NET. Yet instead it has become a detriment.

This one shows a remarkable lack of understanding of the dynamics of a developer community. Open-source frameworks don't spring up, pardon the pun, because there's no central owner to the language.
They exist because people need more functionality and the language designers originally built it to be easily extensible. The reason there are many more Java frameworks out there is simply because Java has been around longer, enough to have several good frameworks mature and spread. Many of those are being ported to .NET and others will rise up.

Original Article wrote:

And even there, Visual Studio .NET only integrates with Visual SourceSafe, another Microsoft product, rather than the all the other choices like ClearCase, Subversion, etc.

Simply not true, incidently, but we'll chalk this one up to lack of familiarity with the tools.

Original Article wrote:

In the area of other tools related to agile methodologies, only after some time did the fledgling open-source community on the .NET side create equivalents—NUnit, NAnt, and CruiseControl .NET—to their Java counterparts.

Quite true. .NET has been lagging behind Java when it comes to tool support, just like it is with application frameworks. This is to be expected from a newer player in the field. But as he mentions, you now have NUnit, NAnt, CruiseControl.NET, as well as MSBuild, TestDriven.NET, Resharper, VisualAssist and many other tools useful for agile development. The gap has narrowed.

Original Article wrote:

Agile methodologies transcend implementation technology, yet Microsoft seems to not be quite on board. Only the most powerful, and most expensive, edition of Visual Studio .NET 2005, has NUnit integrated. No such luck with NAnt.

While Visual Studio has MSBuild and the testing framework built in, you can still use external tools like NUnit and NAnt with it. Why is he insisting on using out-of-the-box tools now, when he was using external tools in his Java IDE?


To summarize my thoughts, I think this article suffers from one basic flaw - he has extensive experience with Java and its frameworks, but has only had a cursory glance at the .NET world and accompanying tools and frameworks. If he delves deeper he will see that the two are more comparable than he thinks. .NET had many missing parts that the Java world has filled in its years of existence, but these are being filled and completed even as we speak. Moreover, these are usually fixed in ways compatible with the original Java implementations - moving from JUnit to NUnit or from Spring to Spring.NET (or the other way around) is probably easier than from totally disparate implementations.

In some areas, .NET is still lagging behind. COM+ is the only application server I know of that can run .NET components, and it's often overkill. A more flexible solution for that in the .NET world would be great. The .NET solution biosphere is still not as mature as Java, but what it does, it does very well - often better than the original Java product.

I think Java is really simple, and C# is some ugly anyway.