"We are using an assumed technology of design all the time, but Alexander argues that what is needed must be different in five key respects:
1) Adaptive design cannot start from a supposed tabula rasa condition, but will always transform what already exists. Even similar design problems, in different contexts, have the task of transforming distinct configurations. In mathematical terms, every design problem has distinct initial conditions that strongly influence the solution.
2) Adaptive design has to engage multiple actors, forming a “collective intelligence” to explore the universe of available solutions and non-solutions. Otherwise, the search algorithm seeking good solutions can take forever, so someone chooses an arbitrary, poorly adapted or dysfunctional solution out of desperation.
3) Adaptive design explicitly employs simple stepwise procedures, operating sometimes at fine scales that can vary and adapt as they develop. This is known in the software community as “interactive computation”, in which the momentary configuration influences the solution as it develops. Computation is affected by feedback in real time.
4) An intelligent approach to design recapitulates the evolutionary successes of the past, and avoids the evolutionary failures of the past, by retaining “genetic information” on the most successful patterns, which we can re-use. Again, it is the software people who have profited most from this insight.
5) A revolutionary aspect is to use the qualitative aspects of living systems, and in particular, the qualities of feeling that we bring to the design process. Surprisingly, this qualitative “selection by systemic attributes” very effectively helps to narrow down the search for adaptive solutions."