The most vocal sources of information on Java-to-.NET comparisions are, unfortunately, the combatants themselves. Those with one foot in the Java world and the other in the .NET camp are actually the people you want to hear from.

I was on a well-known website for the Java community the other day. This site allows people to post news and commentary from around the web in real time. After a post was made highlighting news about .NET, a comment was appended, and I quote: “ Can we stop these .NET posts? This is a Java community web site, not a .NET community web site.”

Ah. There’s the rub. It’s religion to some people. Ditto for the .NET side of the debate. The reality is actually down the middle: business people DO NOT CARE about the ideological aspects of .NET over Java. They simply want something that works.

We’ve been hearing from many customers that are struggling with deciding what to do about platforms. The essence of the struggle goes something like this: VB6 shops have to choose a new platform for new development, because developing in VB6 gets less and less attractive with each passing day. The product will lose support from Microsoft after awhile, and no new revs will be forthcoming. So, the logic goes, we need to look at .NET and Java/J2EE—and DECIDE. Since they both will cost about the same to migrate to, we can choose one or the other for about the same cost.

What should we do? The answer is pretty simple. But it pretty much depends on who you ask.

If you ask a service provider with expertise in one platform or the other, obviously, what you are going to get is a self-serving answer that puts that service provider in position to derive revenue from your ‘choice’. Ditto if you ask Microsoft, or Sun, or IBM. The answers will be predictably biased. So where do you go to get a source of unbiased opinion and guidance on platform selection?

One place to go is a towards a service provider who does consulting and/or training in BOTH platforms. A service provider like this can demystify the comparision, and add value by detailing both the features and the warts on both offerings. Getting a compare-and-contrast from an independent source that has worked with both platforms provides candor that can be both refreshing and truly informative.

Every choice has tradeoffs. While .NET has exploited some obvious gaps in the Java platform, such as relatively crude IDE tools and utilities, there is always a price to pay for convenience. One of the things NewTech instructors are fond of pointing out this truism: “If the tool is doing something FOR you, it’s also doing something TO you.” In the case of choosing Microsoft, the payment is vendor-lock in. For many organizations, it simply does not matter because they are ALREADY a Microsoft shop. Others wary of vendor lock in risks need to balance that fear against the relatively slow-moving Java Community Process (JCP), a protocol for action that includes all the vendors with a stake in changes to Java. This slow moving deliberative process is a negative if you want to move fast. However, the JCP also makes sure that no one pulls the rug out from under you with changes that force you to “rip and replace” your costly code base.

The best way to approach the platform decision is to prioritize your issues, ranking them from high to low importance. Then assign a weighting to each item and tally the score. The topics for consideration should include at least the following items:

1. Execution performance- the resulting app must run well in a production setting, and the execution speed must be verifyable before you start
2. Availability of programmers- Labor for development must be plentiful and easily replaced
3. Existence of developer community to serve as developer resources- developers need places to go to get answers. A real developer community is necessary to get answers
4. Training of developers- Developers need training, or they already have it. How much does ‘fully trained’ expertise cost, when shopping for fulltime and consulting help in developing to the target platform?
5. Documented best practices- The target platform must have documented large-scale apps in production as case studies to serve as guides and benchmarks for future development
6. Usability of IDE tools- The target platform must feature a rich toolset with which developers of average skill can readily learn and use
7. Single vendor lock-in vs. multi-vendor community—The target platform must be carefully examined to determine the level of vendor dependence created. Also, the vendor must be scored against past performance in the area of migration ease, i.e., the risk of having to ‘rip and replace’ due to a vendor making non-backward-compatible ‘upgrades’ to the platform
8. Viabilility of platform vendor—the platform vendor must have staying power and be a force in the industry years from now
9. Maintainability of resulting code—the code must be easily changed in the future
10. Licensing Costs – the costs for all tools, server-side platform licensing and ongoing annual costs must be measured and quantified
11. Likely lifespan of resulting code—the likely longevity of the resulting application code must be estimated based on observable vendor and industry behavior

As you can see, the question “what platform?” actually has many answers….and it all depends. We have recently presented a course outlining these considerations to many interested clients in the past few months.
 
(see Technology Overview: Development Platforms and Tools)
 
In this course we bring our balanced expertise and opinion regarding Java and .NET to the classroom. In this class we ask customers to rank the key considerations and then we supply the score for each platform on each key item. The results are pretty interesting, because in the end, there is no absolutely right answer.

And that’s because each company assigns a unique weighting to each key platform-choice consideration.

We’ve developed a proprietary scorecard that helps companies weigh these choices rationally and make good platform choice decisions. Call us to learn more at (203) 239-6874 or click here to send us an e:mail.

Dan Mezick is President of New Technology Solutions
(North Haven, CT)