Abstract

If building something needs more than one person involved, there needs to be a way they can talk to each other about what shape is to be built. Nowadays we have to talk to software for graphic rendering of images, analysis of performance and automated manufacture, and the language we use to describe shape really has to be unambiguous.

The talk takes the example of smooth shapes ( examples include boats, cars and aircraft) and outlines the methods for defining them used over the last century, how they developed with the increasing amounts of computer power available, and the reasons why some representations have superseded others and become accepted as standards.

Malcolm Sabin developed one of the first surface representation systems using parametric surfaces when at Vickers Armstrongs (Aircraft) Ltd. in the late 1960’s and early 1970’s. He then spent half a year at the Hungarian Academy of Sciences and three years at Kongsberg Data Systems writing code inside a controller for numerically controlled machine tools. He was then head-hunted to support a panel advising Department of Industry and the Science and Engineering Research Council on funding of Computer Aided Engineering. The DoI connection disappeared after three years, but he continued advising the SERC through its ACME directorate until the mid 90s. He has worked with FEGS Ltd. on pre- and post-processing of finite element meshes. In the mid 90’s he set up Numerical Geometry Ltd. doing little bits of consulting, but has also had close links with first DAMTP (Mike Powell’s numerical analysis group) , the Computer Lab (Neil Dodgson’s Rainbow Group) and now the Engineering Department (Fehmi Cirak’s group). He was awarded the Bezier Award in 2010 and the IMA gold medal in 2012 for his work in shape representations.