A Design Framework from Traditional Pattern. Case Study: Batak Toba House, Indonesia

Abstract

This thesis investigates a traditional house design framework by identifying patterns and prescribing rules composed through fieldwork documentation, including interviews, on-site measurement, and observation of traditional houses and their extensions, representing contemporary forms of adaptation.

The case study of this thesis is Batak Toba houses in four North Sumatra, Indonesia hamlets. These culturally valued dwellings are still primarily inhabited, and the region is undergoing intense modern development.

Traditional Batak Toba houses' analysis of this thesis involves computational design analysis, using techniques derived from graph theory to investigate patterns. The study uses justified plan graphs of functionally programmed space as part of space syntax analytics to elucidate intangible social structure and cultural hierarchy within traditional houses. Shape grammar analysis operates the patterns into rule sets, exemplifying traditional houses' transformation and extension logic, with their iteration highlighting a hidden pattern. Computational methods can reveal cultural characteristics that are not immediately evident from visual descriptions or encounters with traditional houses.

The study has found that traditional communities tend to preserve their culturally specific values and way of life expressed in their houses and that adapting to changes in lifestyle through spatial adaptation is a more nuanced process than initially thought. These cultural resiliencies are apparent in the case study’s spatial configuration preservation, variation, and alteration process.

This thesis enriches the knowledge of vernacular design through computational analysis and on-site user engagement. It offers an analytical and generative framework for heritage conservation and future development, using shape grammar and space syntax diagrams as generative design methodologies to avoid the reproduction of eclectic traditional images.