This chapter kicks off with Tompkins talking with Kenoros. Kenoros told him that he had graded the design of every team, the quality of the design didn’t matter to him, what mattered was that there was a design at all. All A teams had an awful score. The reason was that these teams were overstaffed and design was not suited to a large group. Design was better done with a team of 5 or 6 people. When you have 20 or more people, you can’t just have do nothing, so the teams jumped straight to coding. With such a tight deadline, the teams needed to do a last minute implementation without bugs, but with no design that was going to be difficult. Skipping design can lead to some big consequences like the project deviating from its goal or reaching the goal in a non-convenient way.
You can try and divide a project when doing design, but most of the time that results in no design being done. Dividing a project at design time will result in a lot of interfaces between people which in return increase meeting times, interdependence and frustration.
Later, a manager from a smaller team came requesting more people to work under her. This was unexpected at first, since the reason why the team was doing so ewell was because of the small team size. However, they were mostly done with the design and when the implementation part came along, there would actually be work to do for 35 more people. Turns out that the optimal team size varies throughout the project’s lifespan.
“When the detailed, low-level modular design is done, opportunities for splitting up the work explode.”
Projects started with an aggressive deadline tend to take longer than those started with a reasonable deadline. Even if the deadline is set on a later date, the project will probably finish first.
Miguel's Blog 