Response to 5 easy ways to fail

Joel Spolsky wrote an article on 5 easy ways for a software project to fail and I think I disagree with Joels first point. I think in every team i have worked in there has been a star or two, a pile of middle ground developers and some duds. In my experience its more important to get rid of the duds than to get the 'superstars'.

While stars can be great, they are few and far between and they may not fit your corporate culture. Stars also have a habit of doing things their own way which no one else can follow and make themselves indispensable.

Back at HIH a now defunct Australian insurance company, i was working in Oracle Forms 2.0 with some extra additions that had been added by the local guru. They were great and made our developers lives easier, but then we moved to Forms 3.0 and voila these proprietary extensions didn't upgrade nicely. Which in the end created a headache and a lot of extra work in the upgrade process. My point being that the local guru, I don't know his name, did things which probably cost the company a lot of money in the end. I don't really know the politics of why the extensions got added.

I find a true superstar is not just a good coder but a good teacher as well. So people can follow what has gone before. I wonder if thats what Joel looks for when he chooses coders.

But i agree to avoid 'duds' just read any article from http://worsethanfailure.com/ to see the value of avoiding duds.