The subject of this paper looks into designing high density residential architecture that enables certain qualities, in order to improve the livability and perception of the design. A parametric design approach was chosen for this project in order to enable these qualities. First, several important conditions and criteria for this project were formed. These form guiding principles for the parametric models. This is important, since parametric modeling software can often lead to reflexive design criteria and unnecessary complexity. In order to avoid this, relation to context and social relevance become important guiding criteria. View, daylight, orientation and privacy are all guiding criteria that contribute to this social relevance.
The first Grasshopper model was then created. By switching between different domains (or design aspects) within a design process, and combining their results, a qualitative design can eventually be created. The first Grasshopper model, however, forms a more linear process, in which every step within the created model is placed within one continuous code. A second Grasshopper model was then created. This model showed promise by allowing better differentiation between de different design domains. However, by basing the input parameters of the second Grasshopper model on a gradient image, the input parameters became in a way fixed within this model. With the help of parametric form-finding, the architect can create a certain kind of differentiation of elements within his design, giving him more freedom in the composition of the design within the design process. However, parametric codes are created with a certain algorithmic logic. This can be limiting the creative freedom of the designer and a design could then possibly lose its human-touch. Therefore, the designer should always be in control over the parametric code he created, changing the role of the designer from a physical designer to a more virtual designer, a programmer, who is in control of all the aspects of parametric design (input, algorithmic process and output).
By reflecting on each Grasshopper model, in order to understand the strengths and weaknesses, the complexity of the design process of designing high density housing itself could be made more understandable and rationalized. However, according to certain Literature (Simon 1973; Lawson 1980) the design process can never be fully rationalized. This is because the design process can also be described as ill-structured. Understanding the advantages of the computer in well-structured problems and its disadvantages in ill-structured problems, is also important for understanding the need for heuristics within the parametric design approach. Rowe (1982) states that the architectural design process is inherent heuristic. This means that the further a design process advances, more conditions become clear and therefore, the better the design process could be rationalized by the designer or architect. Within the design and development of the parametric tool itself, this would mean that perhaps a parametric code / program could be designed in such a way, that it can rewrite itself according to what it has learned. In languages that support Object Oriented Programming (OOP) an object can be seen as a structure that contains both data and procedures. Understanding these procedures will enable the designer to reflect on the input and output data, making a cooperative and iterative trial-and-error process possible between the computer and the designer.