Environment and Planning B: Urban Analytics and City Science

This article introduces, and presents in a slightly edited translation, a paper written in 1928 by the Moscow architectural student Nikolai Krasil'nikov. Moisei Ginsburg, intellectual leader of the Constructivist group OSA (of which Krasil'nikov was a member), had built the course in architectural theory that he gave to the Vkhutemas diploma class around the procedures of his group's ‘functional method’, which has considerable historical interest in itself. It was, however, Krasil'nikov who saw the critical shortcoming of the method so far developed; he points out that, while it created a firm basis for the introduction of scientific, and above all, quantified methods into the analytical stages of the design process, it offered as yet no means of dealing with the crucial synthetic stage with a similar certainty. In this paper and in a later collaborative one, also quoted, he tried to ‘creep towards’ such procedures, which have interest as precursors of recent trends.

This paper, the third of a sequence, provides illustrations of the theoretical concepts which have been introduced in the previous two papers. Each example points to the need for a hierarchical collection of sets X, X‘, X“, …, where for example X’ is a cover of X. At any one level, say the N-level, a structural backcloth S(N) is obtained, together with some relevant pattern {π}. Changes {δπ} in these patterns correspond to structural t-forces, and these are experienced by the specific structures defined by the data. The first example is essentially architectural and is based on simple data collected for the village of Lavenham (West Suffolk, England). An interpretation of the Q-analysis is offered at both N and (N- 1)-levels. The second example is based on data collected for a study area in Southend-on-Sea (Essex, England). These data are relevant both to commerce and to public amenities, and the Q-analysis provides a means of studying the changes produced over the period of time, 1910–1972, at the N-level. This enables the appreciation of the relative nature of the static backcloth S(N) over a time period and the interpretation of changes in S(N) in terms of structural t-forces which are manifest during this time. Specific graded patterns are produced that are related to car parking and its effect on the intuitive experience of the town structure. The third example is based on a published community study in the University of Essex (Atkin, 1974c). In this case a span of five hierarchical levels is used, and the examples are taken from the backcloths S(N), S(N- 1), and S(N+ 1). These include an analysis of housing accommodation [together with an illustration of the algebraic pattern generator Θ(K)], catering facilities, and the political structure of the university committees. The most striking feature of this study is the discovery of q-holes in the structure, and their roles as q-objects for traffic. In the case of the committees of the community, a new and disturbing light is thrown on the well-intentioned efforts to make the university more democratic and suggests that, without a deep analysis of the structure, these efforts can easily result in the generation of noise as a substitute for genuine participation.

In recent work (Tabor, 1970; 1975; March and Steadman, 1971), spectra of building forms have been studied in terms of their circulation characteristics with respect to occupancy by single homogeneous organizations. In this paper the more realistic assumption is made that a building may well accommodate more than one organization, or a heterogeneous organization made up of departments with high within-departmert interaction and negligible between-department interaction measured in terms of trips. The paper is in five sections. Section 1 outlines an approximate permutational algorithm for arranging the departments within the building forms. Section 2 sets out seventeen distinct but related building forms and defines the circulation measurements to be examined. Section 3 describes the eleven schematic circulation patterns chosen for the experiment. Section 4 looks at the theoretical match between circulation patterns and building forms, finally section 5 employs the approximate permutational algorithm to compare trip distributions over the possible combinations of building forms and schematic circulation patterns.

This paper describes a methodological approach to the design of large residential developments. Techniques are described for predicting the composition of the population in proposed developments, including changes in the age structure and in household formation over time, as the housing development and its inhabitants mature. Forecasts are also made about where the residents are likely to work, what services they will require, and where their recreation time might be spent. It is assumed that the residential units are sufficiently large for statistical methods to be valid in making forecasts. Foreseen variations in housing needs over time are taken into account by the flexibility within the design of buildings. Methods for determining the sizes of auxiliary urban facilities are also outlined. As comprehensive an approach as possible is taken, the methods themselves being subject to improvement with use and experience.

This paper relates generalized measures of information to expected likelihood functions (ELFs) derived from Bayes' equation. It then demonstrates that Jaynes' formalism may be extended to formulate a class of minimally-prejudiced models of which those derived from Shannon's measure are but a limiting and special case. The rôle of probable inference and of information-minimizing models in design is commented on.

The paper follows the preparation of a brief for the design of a school in an English county. Mathematical techniques were made available to help in determining an appropriate schedule of room sizes and to carry out appraisals of alternative proposed outline designs. For various reasons the project was not successful. The author was present, as an independent observer, at the meetings of the working party which was set up to prepare the design brief. The case study is written from a viewpoint which is independent of the viewpoints both of the research unit which developed and presented the mathematical techniques, and of the architect and the local authority which sought to use the techniques.