Graphics standards for three-dimensional modelling

Abstract

Much effort has been spent in defining a standard call interface for a library of routines with display primitives embedded in a three-dimensional space. This paper argues that this is now an inappropriate target. There is too much diversity in the three-dimensional things we want to draw and in the kinds of picture we want to draw of them for any single standard to be both broadly acceptable and specifically useful. If the decision is made to go for a set of specialized standards for wire frame and for rendered graphics for data presentation, for sculptured surfaces and for object models, it becomes apparent that far more can be gained from a higher level standard, at the data presentation or object modelling level. In the object modelling community, at least, there is indeed work toward such a standard based on the IGES proposals. This has very few concepts in common with the graphics standard effort, and it is not clear at present what contribution the graphics community has to make in this area. The suggestion made here is that there may well be a number of paragraphs or sections common to more than one of the graphics-related and higher level standards. One such is the coordinate transformation (the dreaded 4 × 4 matrix). Its definition and its use both for modelling and viewing concepts are suggested which resolve some of the black art associated with transformation in the past. These could become the conceptual basis for a short simple standard in this area.