• 99 consultants from eight companies took part in a one-day experiment that compared two different object-oriented control styles.
• 196 consultants from 23 companies in Norway, Sweden, and the UK took part in a one-day experiment that tested the effects of pair programming.
• 20 programmers from 13 companies worked individually from one to two weeks in an experiment on UML.
• In a study on effort estimation and development processes, 35 companies presented bids for a Web-based system we needed. Four of them were selected to build the system independently of each other. The teams (two or three developers from each company) spent from seven to 25 person-weeks each.
• In another, similar study, 30 companies from 11 countries in Europe and Asia presented their bids. Four companies built the system, each spending from 10 to 20 person-weeks.
• Tylenol: Drug A and
• generic acetaminophen: Drug B.
• management-mandated budgets and schedules without apparent awareness of or regard for what needs to be accomplished or what it reasonably will take to accomplish it;
• considerable requirements and scope creep, regularly resulting in unplanned and generally unaccounted-for post-delivery work necessary to render the project workable; and
• quality generally being sacrificed to deadline, with testing invariably inadequately planned/designed, left until the end, and then squeezed in.
1. schedule (have it ready by the stated date, no arguments),
2. functionality (it obviously must perform the required tax processing, but some of the bells and whistles can be left out for now),
3. quality (bugs in the first release will be accepted as long as there are workarounds), and
4. cost (keeping costs down is nice, but it'll cost what it's going to cost).
1. quality (no bugs whatsoever, if you please),
2. functionality (the software must do certain things, but we can negotiate on some of the bells and whistles),
3. schedule (keeping in step with the rest of the aircraft development is a good idea, but if any software problems exist, it's better to take the time to solve them than have the pilots discover them), and
4. cost (as for Example 1).