Research in computer support to industrial design typically involves much experimentation. Proposals for new modeling methods are commonly evaluated using pilot software implementations. The experimentation tends to be lengthy and therefore expensive. We propose a faster way to evaluate the functionality of design tools. This evaluation method, called abstract prototyping, can be used in a very early stage of the development of a computer tool, prior to any software implementation. The evaluation focuses on how potential users appreciate new modeling functions by simulating their basic properties. The method aims at decreasing the risk of building a useless tool. We have applied abstract prototyping to make a selection among different candidate shape modeling techniques. The method proved to be feasible, fast and inexpensive
