The design drawing is an important medium for establishing design support by means of computers. Architects intensively use graphic representations to communicate their design ideas personally, between professionals, and others. In this study, we consider line drawings such as sketches or drawings. Based on previous investigations, we propose that there exist well-structured graphic representations termed graphic units. Examples of graphic units are: grid, zone, axial system, contour, and element vocabulary. Associated to graphic units are specific kinds of design information that is relevant for this kind of representation (for the grid: module size, modular coordination, dependent grids; for the zone: definition of functional elements, dimensions of zones and margins, etc.) Therefore, understanding graphic units forms a basis for computer-based interpretation of drawings during the early phase of the design process. In this paper we present two principally distinct applications: “paper plus” and “pen plus.” The “paper plus” approach features automated recognition of graphic units as the architect is drawing. Work in this area has been based on techniques from multi-agent systems and Case-Based Reasoning. The “pen plus” approach features drawing tools based on graphic units. Work in this area has been based on techniques from expert systems and computer graphics. The “paper plus” and “pen plus” approaches show how an earlier understanding of graphic representations in architectural design is possible, thus lowering the threshold for the use of computers in the design process.