∂ part of node [[notsof]]


See annotations/highlights in [[hypothesis]] above for inline commentary.


Part One

Goodness of Fit

  • The ultimate [[object of design]] is [[form]].
  • [[form]] + [[context]] == [[ensemble]]
  • It is peculiar how much of the comments on [[ensembles]] apply to [[systems]] in general. This book so far often reads like [[christopher alexander]] has independently discovered principles in use in [[computer science]]; or, of course, perhaps [[computer science]] is just one possible take on the nature of [[complex systems]].
    • I am curious to know if people in the reading group with different backgrounds to mine, but whose primary activity is also not [[architecture]], get similar impressions w.r.t. their disciplines.
  • The [[rightness of form]] depends on the degree to which a form fits an [[ensemble]] .
  • Ensembles can be decomposed into many (nested, overlapped) [[form-context boundaries]].
    • But in any given analysis the boundary should probably be kept constant.
  • [[p19]] "The form is a part of the world over which we have control, and which we decide to shape while leaving the rest of the world as it is. The context is that part of the world which puts [[demands]] on this form; anything in the world that makes demands of the form is context. [[Fitness]] is a [[relation]] of [[mutual acceptability]] between these two. In a problem of design we want to satisfy the mutual demands which the two make on one another. We want to put the context and the form into [[effortless contact]] or [[frictionless coexistence]]."
  • [[p20]] interesting/surprising use of [[impossible]].
  • [[p21]] negative perception of fit, in the sense that we notice [[misfit]] and see its absence as [[fit]].
  • [[p24]] achieving good fit via neutralizing misfits
    • [[simonpjw]] comment about misfits as beautiful
  • [[p25]] an exhaustive description of requirements for good fit probably includes the universe. This problem only seems tractable usually because of shared [[context]].
  • [[p27]] a procedure based on listing misfits "most likely to occur". See comment.

The Source of Good Fit

  • [[p32]] Who is [[macbirdy]]? I like their reading.
  • [[p33]] A lot of this chapter feels a bit problematic, outdated in its description of "unselfconscious cultures".
  • [[37]] [[macbirdy]] on a roll
  • [[meta]] my reading is turning more critical/nitpicky as I go through this chapter.

Meeting with the group

  • Andreas said that it was sometimes hard to remember 'design' meant architectural as the ties to graphic design were so solid
  • [[armengol]] primes and [[factoring]]

The Unselfconscious Process

  • I still find this 'unselfconscious' characterization problematic, but I'll let it go -- if I don't I won't get the point, probably.
  • I liked the [[slovakian shawls]] example.
    • Recognizing and curbing bad mutations != ability to introduce good mutations.

The Selfconscious Process

Part Two

The Program

The Realization of the Program

  • Two phases:
    • analysis, where the right design [[program]] is found
      • a tree of sets of requirements / the previous graph
    • synthesis, where the right [[form]] is derived from the program
  • examples of [[diagrams]]
    • [[le corbusier]]'s [[ville radieuse]] is a diagram stemming from two requirements: high density and equal/max access to sunlight and air
    • a [[sphere]], for the requirement to enclose a maximum volume within a surface
    • engineer sketches, [[kekulé]]'s representation of benzene
    • in general they:
      • summarize aspects of physical structure: a [[form diagram]]
      • summarize functional properties or constraints: a [[requirement diagram]]
      • just one or just the other isn't very useful; a good ("constructive") diagram has elements of both.
  • two ways to describe [[form]]:
  • a unified description is the abstract equivalent of a constructive diagram
  • "the program is a hierarchy of the most significant subsets of M"


  • performance standards are based on scales that estimate fit/misfit in certain domains. quantifiable; miss the [[qualitative]].
  • we model problems with fit/misfit as booleans; although they might actually represent quantitative continuous variables plus a threshold, or qualitative opinions by people explicitly expressed
    • fit = 0
    • misfit = 1
    • "Let us remind ourselves of the fundamental principle. Any state of affairs in the ensemble which derives from the interaction between form and context, and causes stress in the ensemble, is a misfit."
  • [[misfit]] is left as a primitive notion (undefined)
  • [[domain]] is the totaility of possible forms within the cognitive reach of the designer
  • [[probabilities]] enter the stage! Finally. They seem to make sense in this context.
    • as usual, when probabilities are independent all's nice and well.
    • [[pearson correlation coefficient]]
    • some correlations between probabilities are intrinsic, some are circumstantial
    • we call correlations [[causal]] when we understand the rules that account for it; like e.g. thermodynamics as they affect materials
  • p111 [[correlation matrix]]
    • helps us define L and thus settle on the G(M, L)
  • three formal properties of G(M, L)
    • L describes all the interactions between variables there is. Because L is the set of two-variable correlations, M must be chosen so it's free from n-variable correlations with n > 2.
      • superordinate concepts like "economics" and "acoustics" tend to be problematic here, and they should be seen as intermediate steps -- must be broken down
    • the two-variable correlations must be small for any pair.
    • p(xi = 0) should be the same for all i (?)
      • but in the next sentence Alexander moves to "roughly the same in significance", which is both distinct than the previous and more reasonable IMHO.


  • the tension between form-making (integration) and analysis (fragmentation) is resolved by "finding knots": problems are not homogeneous, they actually have some structure, so there are right ways to tackle them.
  • p124 well decomposed sets will tend to be densely connected internally but have few or no links (the latter usually being not achievable) between sets
  • "It is the culmination of the designer's task to make every diagram both a pattern and a unit. As a unit it will fit into the hierarchy of larger components that fall above it; as a [[pattern]] it will specify the hierarchy of smaller com­ponents which it itself is made of."


  • "My main task has been to show that there is a deep and important underlying structural correspondence between the [[pattern]] of a problem and the process of designing a physical [[form]] which answers that problem."