An interesting editorial for sure, but as I read it I was struck that the bigger question in what you are describing may be that this project involved a poorly designed, patched together application more than it involved Specialists versus Generalists.

When I was a much younger developer, one of the best pieces of advice I ever recieved was very simple: "KISS - Keep It Simple Stupid". Now at the tail end of my career I would say that most of the software disasters I have seen have come from two things: Developers running amock trying to be too fancy or complex, and a complete lack of commenting in code.

No two other things combine to make an apocalypse-in-waiting than some developer doing work years ago and trying to nurture job security by making code overly complex, and ensuring that there is no commenting for a future developer to know what they were thinking.

As much as I agree that often Specialists can be more helpful than Generalists, if we are talking about what is in fact a software mess, the question becomes somewhat moot.

This is why today, every week, we do code reviews. Everyone gets together and reviews each other's code. And if you work for me and dont comment your code, you wont be working for me for very long. Software is an asset, and an investment. I have no patience for people who think following some theoretical methodology protects that asset. Eyeballs are what protects the asset - look at the code and do so with an eye toward someone someday having to tinker the code.

Concerns about Specialists versus Generalists come second - because if the software is well, a piece of crap, it really doesnt matter who is going to work on it - its already a disaster.