Project Gestation and Team Size

Software projects need time to mature. A number of times I've estimated a project to take, for example, two months. A manager then steps up and says great, we'll stick four developers on it and we'll have it in two weeks time, right?

No.

1. You can't have four people designing and specifying the same project at the same time. Yes, even if it's a really big project four architects is way too many.

2. Twice the developers does not halve the development time. There's an overhead when people work together; they need time to talk to each other, time to integrate their work and time to shoot the breeze. Sadly, 1 + 1 != 2, even in a good team it's probably closer to 1.5.

3. Testing: Finding and fixing bugs takes time. Often you can't find bug B until you've fixed bug A. Doubling the number of testers often means that you'll just have twice as many people finding the same bugs.

I guess what I'm trying to say is that you can cook a pie quicker by turning up the heat, but that might not be a good idea.

On the other hand, I think having one person working alone (on all but the smallest and most unimportant of projects) is a huge mistake. I hate working alone. I can't run my ideas by somebody else who knows the project inside out and the person who reviews my code doesn't have any stake or interest in the project. People working alone on a project feel isolated and miserable. It's no fun.

Finally, when I get bored and leave, you'll have nobody else who has a clue about what I really did and why.