Abstract: Some algorithms for pretreating the input data of orthographic views in 3D solid reconstruction are presented. For views being drawn in the same graphic layer, auxiliary lines parallel to the coordinate axes are added and progressively relocated until the three orthographic views are separated from each other. Since the graphic entities are originally presented in the device coordinate system, transformation matrices relative to the new origin of separated views are needed to transform each entity to the new local coordinate system. Tangent point and crossing point between line segments are referred to as implicit points. They should be explicitly marked out to help the tracing of wireframe loops. All the classified points and line segments are stored in the corresponding chained lists.