is an excellent book by Don Reinertsen (isbn 0-684-83991-1). As usual I'm going to quote from a few pages:
Best practices are only "best" in certain contexts and to achieve certain objectives. A change in either the context or the objective can quickly transform a "best practice" into a stupid approach.
We must constantly remember that the simplification [of the abstraction] is a veil between us and the reality of the process.
Manufacturing is an inherently repetitive process and therefore more likely to produce learning than a non-repetitive process, like product development.
In the design process we can only add value when we do something differently. If we change nothing, then we add no value.
We will see much more variability in the design process than we see in repetitive manufacturing process. This variability is an indicator that we are doing something new in the design process, and thus that we are adding value.
Risk and the variability associated with it, are inherent and desirable characteristics of design processes. They are at the heart of a design process's capacity to generate information.
The design factory needs variability, so it needs tools that allow us to coexist with variability.
When you dig below the surface you discover that most change programs are motivated by two forces: imitation and obedience.
Events that are less probable contain more information.
Our testing processes need to have an adequate failure rate to generate sufficient information.
Each iteration starts on the base of the quality of the previous pass.
A process is a way of preserving learning that occurs when doing an activity.