INTRODUCTION

During the design process, architects and urban designers are constantly seeking to predict how users will behave and what the character of the projected space will be. Even guided by experience, sometimes designers fail, leading to unexpected situations, sometimes acceptable while at other times undesirable.  Nowadays, it is already possible to simulate the performance of a built environment according to structural, environmental or energetic parameters (LAMBERT and HENSENAND, 2011). However, it there still is a lack in relation to the ways in which users will develop activities in a building. Simeone and Kalay’s (2012) event-based model to simulate human behaviour was one of the first attempts to develop a system capable of simulating human activity, going beyond crowd or flow simulations (BRAUN et. al, 2003). In indoor environments, events are usually a result of human activity, but in outdoor spaces, events can be a result of the interaction between other dynamic objects, such as wind, sun, clouds, water, etc. These events and the spatial arrangement between objects assist in giving the character of an outdoor space, which in turn can attract or repel users. Thousands of events recurrently and concurrently happen in outdoor spaces, making it impractical to represent all events of the real world and attempt to predict the character of a place. As a result, it is necessary to look for the more meaningful structures in space, which are singular, and have a greater influence in the user’s decision-making to choose a place to develop an activity.

In 1976, Christopher Alexander, an Austrian mathematician and architect, published A timeless Way of Building. This book represents the starting point of Alexander’s research to understand why some structures built by men are more ‘alive’ than others. Looking at medieval cities, the author perceived that what matters in a place and gives it character is not only its physical geometry, but also the human and non-human events that happen there. Alexander also observed that in these urban centres, citizens built spaces over and over again according to sets of informally defined rules, generating similar design solutions adapted to the context. In the following years, the author began to extract the rules from similar structures built by man, creating patterns organized as a generative language.

In The Nature of Order, Alexander (2002) extend his theory of patterns, looking for structures that can assist in explaining environmental wholeness. According to him, every structure has a set of centres that is responsible for defining its degree of ‘life’. These entities are a distinct physical set, which occupy a certain volume, and have a special coherence in space. The author explain that the centres we see, when we look for wholeness, are not the centres which are captured by words, such as “bike”, “road” and “tree”, but a set of centres which are not described by single words, and are induced structurally, by the overall configuration of a scene. Alexander concludes that when a building or place is a ‘living’ structure, centres will be strong entities, and recursively interconnected with each other. The result is a built environment that makes us feel comfortable and generates a sense of belonging.

The purpose of the current paper is to introduce the preliminary results of research that seeks to develop a theoretical model to describe the character of outdoor space and structures responsible to attract people, using as a reference point Alexander’s theory of living structures. The proposed general hypothesis is that an adequate description of a place can lead to simulations capable of predicting the dynamic and performance of an outdoor space and assist landscape designers to comprehend the impact of their decision-making during the design process.

METHODOLOGICAL APPROACH

The research was developed in three different phases. In the first phase, a diagram was built employing as reference Alexander’s theory of living structure. In the second one, a lawn in a university was selected as case study and naturalistic observation was conducted at the research site to collect data of people and their behavior in this specific environment (SCHENSUL, 1999). Finally, a set of rules was generated using shape grammars (STINY, 2006) and object oriented first order logic (MITCHELL, 1990) to express the interaction between physical objects, non-human and human events.

Figure 1 shows the diagram developed during the first phase of research. It aims to demonstrate that the interaction between different objects, more or less dynamic, is responsible for defining attractiveness and that, as with any other object, it can also be instantiated, emerging or not in the place depending on contextual factors. For instance, we can define the wind that moves the leaves of a tree as one object. If the attribute’s velocity is low, it will generate a breeze that will gently shake the leaves and this movement will produce a nice sound. These events are the result of interaction between different objects and are responsible for giving a quality to space, providing an environment that attracts people to do a certain activity. In the other hand, when the wind velocity is high, not only the leaves, but also the branches will move and the space maybe will not be comfortable. In this case, the quality is latent and it will emerge again when the attribute’s velocity has suitable values again. However, if we replace the object tree by a built roof this will lead to a different situation. The attractiveness generated by the sound and movement will not be there anymore, because the new object has its own attributes, leading to different interaction with the wind. As a result, maybe this place will be less attractive than the one that has the tree.

As previously mentioned, systematic naturalistic observation was applied in the second phase as a methodological approach to gather data of how people behave in space. The goal was to observe what environmental aspects could potentially attract or repel users developing activity at the research site. The place chosen to be studied was a grass lawn in front of the Computer Science Faculty because this is a well-defined space, with clear borders, surrounded by buildings, a rectilinear path, a water mirror and some bushes and trees. The site was observed over one month in November 2014, from 08:00 to 17:00. Sketches of peoples’ disposition and formation, annotations and images were collected during this period and retrospectively analysed to seek for patterns of people distribution in space, intensity of movement, and body formation.

In the third and last phase, drawing on the information collected in the previous phase, we decided to focus on the volumetric space around trees because preliminary conclusions from observation indicated that these were preferred spots, much more dynamic and where multiple activities recurrently happen.  In this phase, using as reference shape grammar’s rules and first order logic sentences, it was possible to express the general dynamic of the lawn (the character) and the selected spaces around trees (the atractive spaces or the centers).

RESULTS

In the chosen space, people are attracted to do an activity around a tree not only because of the tree itself, or the grass, but also by the wholeness generated by the interaction between a set of objects. For instance, the shadow around the tree is dynamic, does not have a clear transition to direct light and is instantiated during the day, according to the presence or absence of clouds covering the sun. The shadow has only these attributes because of the interaction between different objects and its presence contributes to generating a volumetric space adapted to receive users in more different conditions. For instance, in a cold sunny day a user can dispose the body in the sun and read a book in the shadow, while in a hot sunny day a user can seat completely under the tree. If the person is using a computer, the tree shadow provide spaces where the light is partially filtered, however a built rectilinear roof usually do not provide a transition between direct and indirect light. As will be shown in this paper, the volumetric space where objects are immersed can be represented as a hidden object that inherits attributes from surrounding dynamic or static instances.

DISCUSSION

In computer aided design and building information modelling tools, the real world is usually represented by a combination of parametric objects that can, in some systems, keep geometric relations between different instances. Mainly, these systems are developed to generate detailed geometric and visual representations of the real world, however the set of volumetric spaces surrounded by the model, generally do not inherit any attribute from surrounding instances, being treated as void virtual space. Acknowledging that these voids are also objects that attract or repeal users according to surrounding dynamics or static objects can help us develop modelling or simulation tools focused not only on technical or visual aspects, but allowing us to use the full suite of variables potentially present in a space to predict human behaviour in outdoor spaces.