1 comments
Various 
Thanks to my coworker Ronald who passed around this link to a research document on IDE Productivity on our internal Java distribution list. As my company does both a lot of Microsoft and Java related projects why not make it available to a much broader public. The research paper is stated as a comparison of various productivity efficiencies and techniques between the IBM® J2EE and Microsoft® .NET development environments, through the creation of a real life web-based application. Research was conducted by the Branhan Group which are an independent consulting firm from Canada. You should know up front that the assignment was paid for by IBM. This might taint the outcome a little bit of course, but for what its worth the outcome does largely match my opinion on the subject.
Although Microsoft Visual Studio is generally believed to be trendsetting in IDEs it is no longer ruling this domain per se. In the past few years some of the Java IDEs, especially IBMs have had huge improvements in both useability and features. Together with the improvements and extensions in the Java language itself it is no longer the truth to just claim that Microsoft Visual Studio is a more productive environment compared to all current Java environments. If you believe in the outcome of the conducted research it is even better to claim it the other way around. Seeing is believing so you might want to take the new Rational Application Developer 6 or just Eclipse 3.1 for a spin to see what I mean here. Also, take note that Microsoft was in fact present at the recently held EclipseCon 2005. Rumours around the internet even have it that they are considering Eclipse as a lightweight alternative to Visual Studio somewhere in the future, but I really don’t know if this is the truth or just some wishful thinking :o)
As my Microsoft-minded coworkers like Marcel and Raimond are heavily investigating Visual Studio Team System, which is the next-gen Visual Studio, it would be very nice to have another comparison between VSTS and the next-gen Java environments somewhere soon.
One comment
While this is an interesting report, I can’t help thinking that the problem statement and requirements were specifically crafted to get to the weak spots of the Microsoft development tools.
Also hand coding test apps etc would not happen in the real world, as people would just use NUnit as a testing harness. (But no addtions were allowed outside the vendor tools)
Also, the decision to use Services Without Components is strange, as there is complaint about not have caching, connection pooling etc available which would be there with Serviced Components…
Raimond