We suck on estimating size
Still the unsolved problem in software development
So, as a contracter:
Sources and further readings:
- Would you rather pick someone who promises they are right about the scope (.. lying..) and pay their risk premium
- .. contract them, negotiate away the risk premium, only to see skilled resources shift to more profitable projects (once you are firmly in their grip)
- .. contract them, and see them default
- .. Or - pick someone who is honest about this fact, but promises to deliver runnable software monthly with what you require
- Would you sign up for a contract which you know is bad
- Or pick a better client which you can have a thriving business relation with?
Sources and further readings:
- Scott Ambler, Dr Dobbs Journal, "Is fixed-priced software unethical"
- The Standish Group's chaos report
- "Stop estimating" - David Anderson
- Software Cost Estimation with Cocomo II, Addison-Wesly 2000
Re: We suck on estimating size
Many times "deliver working software" is not enough. Estimation tries to answer questions such as, "will we meet the marketing window?".
The funny thing about that is that we tend to forget the question when we start estimating. Instead of focusing on the market window, we focus on what happens to be the set of known requirements.



