I would really like to see that article. Because, I too have been involved in such things, several times, but have not managed to identify exactly where we went wrong (or even convince others that we were going wrong :) ). I think figuring out how to avoid that kind of complexity is one of they key challenges of certain kinds of software these days.

(For instance, some people think Rails has already jumped that shark, some people don't, but I'm not sure even the people who think it has become complectified all agree on when it happened and how it could have been avoided. When I see people starting something new that's supposed to be "like Rails, but without all that needless complexity", I just think, sure, Rails started out simpler than today's Rails too, and you'll end up in the same place (or worse) unless you can come up with an understanding of what went wrong other than "Those other people made bad decisions I would never have made because I'm a better coder", nope, that's not what happened.).