I just came across these, and thought they were too important to just put on the linklog. They deserve to be the center of attention.
So here they are, the 5 pitfalls of estimating a software project with my comments:
1) Allowing non-technical staffers to give estimates.
Aaaarrrgghhh!! That so gets on my nerves when someone non-technical just goes ahead and gives a baseless estimation.
2) Being afraid to look in the mirror.
Not evaluating past projects and finding out what went wrong in them is one of the biggest mistakes anyone can do. It’s like doing the same thing over and over again and expecting a better outcome.
3) Underestimating design time and debugging time.
This is another major problem. People just love to assume they’re perfect and that they can just skip design and just jump directly into development and then produce something that doesn’t even need testing. Simply stupid.
4) Inadequate/unclear requirements.
Don’t get me going on that one.
5) Taking too large a bite from the apple.
Giving estimations without breaking the tasks down to bits and pieces is just asking for trouble. The complexity of software lies in it’s smallest details not in the big picture.
[Via: Christopher Hawkins]