Cool!

serializing an xml document is not a cost-effective procedure. that is to say, the application is too complex and subjects to a plethora of bugs

Dear Comics, thanks for the comment!

In no way we're suggesting that jxom/C# xom is the most efficient way to generate java/C# code. However, there are many cases where xml/xml schemas/xslt fit very well.

In fact whenever your data is tree structured and you need to manipulate it, and in particular to produce java/C# code, then jxom/C# xom could be a good choice.

jxom now is in a stable state, and C# xom will achieve the same stability soon. Both are used in the production projects.

Comics,

In fact, the cost-effectiveness is subjective measure, and depends on task you're implementing. If you have to create and then, on a later stage to do a code refactoring and optimization then C#/jxom + XSLT are better choice than many other solutions. If you need something for real-time processing then take a look at code-dom for .NET, or other solutions for Java.

How and what tool to use depends on a triplet "subject, tool and task",and affects a general productiveness.