Discussion on:

J2EE vs. .NET: Making that vital development decision

i'd say if you focus on mid-small size customer, you better choice .Net. but if you focus on large company, J2EE is a better technology for you. at least it's true in coming three years.

with tomcat & apache , mysql & jboss ,
all free why are we all locked into the idea that J2ee must be on large solaris / Aix / z/390 boxes, while microsoft is meant for the small shop.

This is a mindset that dominates the industry. Let us get away from it .

A reality check... you have to live with both
My comments are based on real, current and live experience on real projects!

The reality is that you have to use both environments. The secret is not the actual platform, but the quality of the architecture! The SOA (Service Oriented Architecture) is the way to think.

Of course, from real experiences, on multiple projects ranging in size from $50K and $30M; I would pick J2EE anytime. Again, based on past experiences, I have seen much greater rate of failure and unstability on the Microsoft platform.

Bottom line, be wise and don't discard either framework.

One final comment... the last line should have read:
"...but comes down to three: .NET-proprietary, J2EE-generic or J2EE-vendor specific..."

Unless J2EE improves the developing process, Microsoft will of course win! We are developing for IBM Websphere and had to buy a 1.5GB-machine. The edit-test-debug-cycle is still many minutes. This is stone age technology. The productivity of the developers is near zero!

Have you tried IBM's eclipse with Lomboz plug-in for J2EE development?

Why waste time by tip toeing and try to be politically correct-confront the real issues, maybe publish articles written by developers who can voice there views.

My department is currently moving to VB.Net/ASP.Net from VB/ASP. Java was never really a consideration as none of us has much experience it with it. C# also fell out of the running after much discussion at conferences and research on the web seemed to show there is no performance difference between VB.Net and C#.

I believe that Microsoft launched C# to attract Java developers and I suppose this makes some sense, though I don't imagine that a lot of experienced Java developers will be persuaded to change when there is so much well-paid Java work out there. At one conference, a speaker suggested that C# was really created as the internal programming language of Microsoft, which they decided to release to the public. Interesting thought.
It seems to me that the effect of C# is to divide the .Net community, particularly the VB community. Given that there are more VB programmers than any other kind, this seems a strategic error on Microsoft's part. Language adoption depends enormously on the quantity and quality of resources available to programmers to resolve their problems. By dividing the .Net community, Microsoft has divided the resources available to both VB.Net and C#. I see this every day when looking for answers on the web.

This can only strengthen the Java community.

It is my opinion that C# is a new logical step in evolution of new modern-day languages. To really compare and understand nature of C# someone has to know and see at least three different paradigms in programming languages. From one side we have classical C++, from another side we have Java and from the third side we have growing number of scripting languages such as Javascript. Merge all these three paradigms together and you will get C#. The question is, should you learn C#. Let?s assume if you know C++ and Java and Javascript, wouldn?t it be convenient to implement solution in the language that combined power of these three? It is certainly beneficial even from education perspective to take a look at some key language aspects.

To say that the promulgation of C# and VB.NET creates a "split" in the developer community is to misunderstand .NET. Languages that conform to the .NET specifications are essentially created equal, and choice of language is essentially a "lifestyle choice". C# and VB.NET, in terms of the .NET CLR, are essentially equivalent and we can swap one for the other given our personal preference. We can even, for instance, take a component written in C#, decompile it to IL, then recompile it as VB.NET.

It also matters about cost/ performance/ scalability / portability of the system.

Microsoft cost and performance and scalability (only for some microsoft products)

J2EE scalability and portability and performance (only for some vendor specificproducts)

This is under current situations

In the article, its nice to see that they are focusing on two major factors that will be a deciding factor in my company. However, they didn't focus on the fact that a majority of the younger IT professionals out there today are being trianed in Java, and thus using J2E may reduce my overall company cost of retraining staff, and down time due to workshops. In some of the areas here in Pennsylvania, Java is the overall main language for even starting programing courses. Whereas it used to be C,CPP, or, gods forbid anyone else remembers when they taught Fortran as a beginning language; Its now Java. Given this trend, and Sun's ability to capture not only the IT marketplace, but the educational Market place when it comes to turning out future IT pros, I feel the article may have been misslanted. I, and my company in turn, ramin on the fence for now with this issue. I know how to use Java, and so J2E looks more usable. However, I am not above attending a workshop to see how MS's productworks.

Good article. I am glad to see what appears to be a balanced and unbiased opinion about the 2 major IT architectural philosophies that almost all large organizations are having to decide upon, and in some cases decide quickly. In most organizations this decision is shaping up to be made in the midst of "religious wars" between the 2 technical factions that may exist inside these organizations. But let us not forget that the best marketing always wins the IT wars - not the best technologies. I can hardly wait to see who wins.

Microsoft was able to rally industry support for standards such as ODBC because they were entry points into the proprietary Windows OS and Visual Basic development environments. It was a win-win situation. If Oracle and other vendors wanted developers to produce applications using their databases, but still able to use the most popular tools for the predominant desktop platform, they had to adopt ODBC and create an ODBC driver, or do all the hard work themselves. At that time, SQL Server was not yet strong enough to be the only target for Visual Basic; it would have hindered its further adoption. MS and the other vendors realized this and accepted it and felt comfortable with the "open Microsoft standard."

Times have changed, however. Now there is OLE DB and ADO as the "new" standards. Microsoft has their native OLE DB provider for SQL Server, and OLE DB gives it special treatment which gives MS yet another edge. Oracle has developed their own OLE DB provider, but many of the othervendors have been stuck with the overhead of having OLE DB go through ODBC. So once again, MS is managing to use "standards" to enhance their dominant position in the industry. (I admint that there are plenty of other factors and good arguments for the transition, but this still appears to be a major result of the new standard.)

One of the major factors that has caused Unix to lose ground to Microsoft is that the Unix industry is so bitterly divided. While there were "Unix standards" and it was an "open system," there were too many difference between vendors' Unix variants.Just because you were a good Solaris sys admin or programmer, you couldn't automatically handle an AIX or HP/UX system because of the differences. And the vendors often had a "I don't care, as long as he doesn't win" attitude which prevented a unified front against Microsoft.

With J2EE, we have the potential for a similar situation emerging. Sun solidly holds the reigns on the J2EE standard, but has to try and keep it an open platform while still selling Sun products. But you have to have J2EE Application Servers to use J2EE, and each vendor has to compete with the other vendors, and has to come up with proprietary features. So, as the article points out, you have a divided front.

For J2EE to win over .Net in the future, we have to look to the J2EE vendors to "observe gentlemen's rules" amongst themselves, and not warp J2EE into a non-standard.

If the tools allow you generate standard J2EE code, copy it to another vendor's system and reverse engineer into a proprietary tool suite which enhances development productivity, then we get a healthy environment. If we find ourselves stuck with a development environment because there is no migration via a standard, then J2EE will disintegrate into a bunch of niche products. If "runs on J2EE" means the same as "runs on Intel" to developers, it will be time to seriously consider bailing to .Net.

The author has a major point that you missed in your post.

The choice is, go with open J2EE and suffer performance consequences or Improve J2EE performance with vendor specific enhancements and loose portability.

This fact was in one of his recent articles a few months ago.

Bottom line,

generic programming = generic performance

.Net locks you in to the MS platform, but you must decide if the significant speed increase is worth it.

J2EE will diminish into niche divided products if the Unix world and java cannot decide on standards. This is hopefully something Linux may bring to the realm, but unfortunately, you already have Linux zealots and Unix zealots battling each other.
The one thing MS has on its platform are standards, and thats something that shouldn't be ignored. Yes you say windows standards, but windows is in a sense IS a standard. Itanium, Hammer, will all bring the price/cost ratio in favor of powerful inexpensive 64bit wintel servers or Linux intel/AMD servers.

Don't confuse a monopoloy for a standard. Becuase a product such a windows is widely accepted doesn't make it a standard.

There are committees and review boards that take more than market forces into consideration when accepting standards.

You're talking about "de jure" standards, agreed upon by a committee. But in the IT world, there have always been "de facto" standards as well. That is, things that are in actuality a standard because of their wide adoptance.

Even though these de facto standards may have gotten to be so through unsavory means, you ignore them at your own peril. Life is unfair!

It's good to see a fair and un bias look at these technologies, it is so refreshing to read an article like this.

The Author - gets a At-a-boy for a truely well written and balanced and informative article..

KEEP THESE TYPE COMMING ...

Any programmer or CIO today has to balance the time/cost of an overhaul/rewrite versus the time/cost of adding onto what exists given the complexity of the desired features to be added.

A Java-based shop is going to find that adding and integrating added features with Microsoft technology is going to make less sense than sticking with the status quo, and vice versa.

To suggest that a rule can be made that company x should use technology y for project z is just absurd given the complexities involved in x, y, and z.

From my experience, Web Services, VisualStudio.NET and C# are emerging technologies useful for exploring and investigating in terms of the impact on current practices in the 1 - 3 year time frame whereas Java-based standards have had some time to mature and proliferate.

I would be very cautious about hiring a developer who claims to be an expert in Web Services, C# and VisualStudio.NET. The technologies are too new. Likewise I would be very cautious about commiting fully to a Microsoft-only solution. You never know what they're liable to do in the future. (That MS 'stole' Windows from Apple and how they've positioned Passport should be two very large red flags.)

As others have said, an Apache-Tomcat-Linux-MySQL solution can accomplish the same functionality as .NET without huge (if any) licensing fees. If your company has money and time to blow, then go .NET. But if you're like most companies, under time, budget, and human resource constraints, go with what works - Java.

For someone who's never seen a side-by-side comparison you'd be amazed to note that, in terms of the IDE - where the programmer spends his/her time - Apple WebObjects 5 and Macromedia Jrun/ColdFusion/Dreamweaver are both in their 5th version, have moved to J2EE runtime, and are far, far more refined than Microsoft's 1.0 release VisualStudio.

Looks like I started the with a question with 2 choices and ended the article with the same question remaining but now with 3 choices and not much in the form of framework/problem attributes to make choices.

duh ... what a planet do you come from ?

it's not worth any cents !
i must stop wasting time by reading a bunch of crap like this.

you better to change the subject to :
": J2EE vs. .NET: Making that vital development decision FOR MORON"

how to deactivate my member account ???

Using "Microsoft" and "quality software" in the same sentence bothers me!

Quality software does not need to be rebooted daily or weekly, it runs 24x7 for months or years. Quality software does not need weekly security patches. Quality software does not result in the message "Program not responding". It's great when it works but all too often, it doesn't!

Java is years ahead in mindshare and proven products across a large range of platforms. With SAP Oracle IBM pushing Java - this is the first choice of most corporate mission critical apps.
.NET will be there in the periphery in personal apps - for some time.

Earlier I recall there was talk of Microsofts witdrawal of Java support in Internet Explorere as a problem of Y2K proportions. Obviously such opinions are held by people who are not aware of Sun's plugin or Oracle Jinitiator.

This article should be URLed to wwww.microsoft.com ....

If I'm not running Microsoft Windows, I don't think I'll be doing my development using .NET.