Maven's failing is its lack of support for VS

In the end, it is the users who will drive demand for Java or .NET and not the opinions of developers. This is why I think MDA/MDD is so important today and into the future when new VMs arrive.

On the topic of Maven, I was unable to get it working for VS2005, let alone VS2008 and found the whole experience very frustrating. I was exposed to Maven via AndroMDA. Even though I prefer VS, I am still interested in targeting customers who prefer Java, so AndroMDA seemed perfect.

AndroMDA / Maven seems to be a terrible implementation of a fantastic idea. I don't know whether to help tweak Maven/AndroMDA for .NET or drop it altogether and concentrate on implementing the concept for myself using a VS plug-in.

MDA is supposed to allow the same "model" to be executed, debugged, profiled, optimised, built and deployed on differing platforms (PIM/PSM). AndroMDA / Maven achieves this at a very basic level, but the advanced features of the VS IDE are lost.

In the perfect MDA world, the choice between a Java VM and .NET should be a matter of preference for the customer (not the developer). If AndroMDA / Maven was living up to it's promises, this would be the case. The platform specific IDE should only be needed to do the profiling and optimising before a build.

BTW, inheritence from the same code-behind file is possible in VS2005 and above. I use inheritence to define common routines for Web Services and Forms without a problem.