This paper presents and argues for a disciplined method for the development of software and software-related products. The practices suggested are not new; however, the hurdle for increasing software-development productivity is not the absence of practices, but the lack of their application to daily work.

The methods here balance between simplicity and formal rigor in a way that makes adoption easy, but does not compromise necessary formal aspects, such as explicit, measurable quality assurance, and clear entry and exit criteria. The mutual programming approach can make the introduction of formal rigor easier, because at each level the feasibility of more formalism can be actually measured.