We have not attempted to survey the views of creativity held by workers in the knitwear industry, but the issue emerged in discussions with design students at De Montfort University, almost all of whom shared the common erroneous conception of creativity. This supports our other informal observations of views about creativity, which lead us to believe this is probably a majority view, at least among people without a technical education. The students' knitwear design degree course did nothing to challenge this: it paid no attention to goal directed problem solving, either in designing garments or programming knitting machines, or anything else. Instead the students' assignments concentrated almost entirely on the production of aesthetically pleasing design ideas, with the aim of developing the students' fluency in generating ideas for potential garments from any source of inspiration they might come across.

We suspect that this misconception of the nature of creativity may spawn similar attitudes hostile to designers becoming involved in technical development in other fields, if the development of CAD systems encourages a closer integration between the design and technical specification stages of product development. This is an issue meriting serious sociological study, for which we recommend the knitwear industry as a suitable field.

What designers gain and lose from being trained to use CAD systems and from actually using them is an open question meriting serious research. We hypothesise that the limiting effects of stereotyped methods and doing what is technically easy should be reduced by giving designers sufficient mastery to push the CAD systems to their limits, and by training in methods for breaking out of stereotyped patterns. As far as knitwear is concerned, the technological and economic situation limits the possibilities for research to studies of a few individuals, as the cost of training a designer to the level of competence we envisage would be several thousand pounds. Studying computer based knitwear design in a commercial setting is only just beginning to be possible. Our research into techniques for providing effective CAD support for designers is founded on the hypothesis that given good CAD systems, designers will gain much more in producing feasible designs efficiently than they lose in limitations on their creativity.

8. The Potential for Computer Aided Knitwear Design

If knitwear designers are able to work with CAD systems without having their creativity seriously limited, the efficiency of the design and prototyping process could be increased both by the improvement of training and working practices using the existing systems for programming, and by the introduction and effective use of systems giving more support to knitwear designers. In order for designers and technicians to produce knitting machine programs with much less need for iteration and backtracking, the designers would need mastery over the CAD systems as well as a reasonable amount of access to them, plus a good grasp of what stitch structures the company can feasibly produce at a given price point. This would enable them to develop simple designs to an advanced stage on their own, and to reject or modify their most unsuitable designs before consulting their technicians, and make their own simplifications according to their own aesthetic criteria after they talk to the technicians. (Designers learn to do this to some extent already, by accumulating experience slowly and unsystematically.) Improved technical education should have the further effect of enabling the designers to discuss technical problems more easily and efficiently with technicians and make-up staff, permitting a smoother hand-over of their designs. As we note above, the development of knitted garments happens under often intense time pressure: an increase in productivity and a reduction in the development time of an individual garment would be very valuable. Ideally it would release time to devote more effort to perfecting designs and to producing good versions in different sizes. Achieving this situation would involve a renegotiation of the division of work between the designers and the technicians, whose detailed knowledge of how knitting machines work with different yarns would still be indispensable. However some benefit should be obtained from giving designers a much more thorough technical training, even without greater access to CAD systems. A potentially useful analogy comes from the advent of computer literacy and word processors in offices: the new technology is used most effectively where the new role of secretaries has been systematically thought through (Cassell, 1991).

Neither better use of existing technology nor effective use of future CAD systems is likely to happen without significant changes in the knitwear industry. Both depend on knitwear designers gaining both full access to the CAD systems and sufficient training and confidence to have complete mastery over them. This can only happen if companies buy more CAD systems than the absolute minimum number needed by their technicians, which requires managers to see an advantage in investing in the work of their designers. It also requires affordable CAD systems, a radical change from the existing situation. The remaining limitations lie with the designers: knitwear companies cannot afford to invest a lot of resources in training employees they don't expect to keep much more than three years, so the designers need to arrive trained, and they need to gain more than they lose from designing with and for a particular system.