I think analysis tools like this are more an indictment of ceremony code that surrounds a big object-oriented design than a win in programmer productivity. I r

I think analysis tools like this are more an indictment of ceremony code that surrounds a big object-oriented design than a win in programmer productivity. I really do like the idea of taking an existing OO codebase and being able to probe and interact with it(REPL anyone?). I did a stint doing bug support on two big OO software systems a few years ago and it was hard to wrap my head around a huge codebase quickly - the original devs who built the system were long gone and nobody actually knew much about huge pieces of the code. So anything that would have supplemented reading thousands of lines of code, I would have gambled on. It is awesome that these guys did a user study, but I wouldn't bother using a test with 8 programmers "new to Eclipse" as anything but a sniff test to indicate that a bigger user study was warranted. Lastly, the article itself sits in a weird space - hackers who read it are going to be underwhelmed. The lay reader (or programming newbie) is probably going to pick up a bunch of wrong ideas about object oriented development.