CAD Systems and the Division of Labour in Knitwear Design
Claudia Eckert
Department of Computer Studies, Loughborough University of Technology
Martin Stacey
MRC Applied Psychology Unit, Cambridge
Abstract. The design of knitted garments is an activity shared by knitwear designers (who are almost all young and female) and knitting machine technicians (who are almost all male and usually older). The process involves programming knitting machines using CAD systems, which are designed for and used by the technicians. The designers get much less training and access to the CAD systems than they want. This paper examines why this is, and what would be involved in creating a situation where the designers are empowered over the technology of knitwear design. It concludes that the limitations of the technology cause this situation, which is reinforced by economic and attitudinal factors.
Keyword Codes: J.6; K.4.2; K.4.3
Keywords: Computer-Aided Engineering, Social Issues, Organizatorial Impacts
Published in
Women, Work and Computerization: Breaking Old Boundaries - Building New Forms
Edited by A. Adam, J. Emms, E. Green & J. Owen.
Amsterdam, The Netherlands: North-Holland, 1994, pp. 409-422.
Claudia Eckert is now at
Engineering Design Centre, Department of Engineering, University of Cambridge,
Cambridge, UK.
Martin Stacey is now at
School of Computing, De Montfort University, Leicester, UK.
1.Introduction
The profession of knitwear designer is absent from lists of typical female occupations (e.g. Lockwood and Knowles, 1984), but it is almost entirely female; their closest colleagues, knitting machine technicians, are almost entirely male. Both groups are finding computers increasingly unavoidable: programming industrial knitting machines is an essential part of knitwear production, and this is now done almost entirely with purpose-built computer aided design systems. The male group exists to use this complex technology; the female group does not get the access and competence with CAD systems many of its members clearly want. Moreover, one can make a strong prima facie case that empowering knitwear designers over the CAD technology should result in a substantial improvement in the efficiency of the design process, leading to the production of cheaper or better garments.
o Why do knitwear designers have very little access and control over the computer technology in their industry?
o What is involved in creating a situation where knitwear designers have full access and mastery over the computer technology?
o Would creating this situation be cost-effective for the knitwear industry?
This paper is an attempt to answer the first question and explore the second as far as is possible at the present time. The answer to the third depends on future developments in the industry, particularly on the capability and cost of CAD systems, though we describe the major factors determining the answer. Technological, economic and attitudinal factors all act to create a situation where knitwear designers have very little opportunity to use or develop competence with knitting machine CAD systems; we attempt to assess the relative importance of these influences. In discussing what the obstacles are to enabling knitwear designers to use CAD systems, and how they might be overcome, we have chosen not to discuss future developments in computer support for designing. Instead we concentrate on the barriers that exist to effective use of existing systems and to the development of CAD systems better suited to the needs of designers. We also comment in passing on the sharp sex divisions between the different occupations in the knitwear industry, though we have not studied the causes for the sex segregation we observe.
2.Investigating the Potential for CAD Systems for Designing Knitwear
The primary objective of our research is the development of CAD systems to support knitwear design, that is, to support knitwear designers rather than technicians. To state the obvious, effective use of new technology involves providing the right technology for the needs of the industry, and the right industry for the technology, so that it is not wasted because of bad organisation or training, or harmful attitudes. This paper is a by-product of a study of designers and of the knitwear industry by the first author, which was intended primarily to determine the potential users' requirements for an intelligent CAD system. In order to study the design process and the attitudes of designers and technicians towards design, CAD systems, themselves and each other, the first author employed a combination of ethnographic methods.
As we are computer scientists working on ways to support design activities, our sociological observations and analysis are informal and incidental to our study of the design process, and are not situated in any developed theoretical framework. In particular, we use the term attitude in what we take to be its conventional everyday meaning. We are aware that discussing attitudes in this way is risky and unscholarly, in that it is difficult to say anything with confidence about attitudes, and we were not able to work with the rigour expected in both computer science and the social sciences. Nevertheless we find that the social context of CAD system use in the knitwear industry is too important for our technological work for us to ignore it.
The first author conducted structured interviews with designers, and informal interviews with designers and knitting machine technicians working in the British knitwear industry, as well as design students and design teachers at De Montfort University (formerly Leicester Polytechnic), and CAD software developers at three of the four major producers of CAD systems for flat bed knitting machines (Eckert and Murray, 1993). The interviews provided detailed but informal information about working practices at a small sample of knitwear companies, biased towards the largest companies because these are the primary users of new knitting machine technology, plus information about many other companies where designers or technicians had worked previously or had friends. Some further information about commercial working practices comes from the completed questionnaires we have received from the senior designers working at about 20 of about 75 British knitwear companies we were able to identify. In consequence we have confidence that our general statements about employment practices in the industry have very few exceptions; knitwear companies vary much more in the organisational aspects of their working practices.
The first author attended courses in knitwear design at De Montfort University, thus gaining a detailed first hand knowledge of their curricula and teaching methods, as well as information about the attitudes of both students and teachers.
3.The Knitwear Design Process
Knitwear design is a process and a profession distinct from fashion design, which makes use of the output of textile design, and is principally about designing exact shapes. By contrast knitwear design involves the creation of the knitted fabric itself as an integral part of the design of a garment, in which precise shaping is usually less important (because knitted fabric is much more stretchable than woven fabric). (The other major division of clothing design is contour design, which covers underwear and swimwear.)
The workers in the knitwear industry are divided into several very different groups performing different roles, recruited in different ways from different social groups. The development of garments to the stage where they can be mass produced is a collaborative effort between designers, technicians and sampling make-up staff (Eckert and Murray, 1993). In order to explain the use of CAD systems and the roles of the different groups, we describe this process for a typical company large enough to afford CAD systems; there is considerable variation in working practices.
1. Research. This is the term given to absorbing ideas (from other people's garments or the outside world), and learning about the ideas, topics, colours and features "in fashion" for the coming season. In cognitive science terms, this constitutes the construction of a search space for the design of a garment. The degree of scope the designer has varies greatly; sometimes the designer's brief is to copy someone else's garment making the minimum number of changes to evade the copyright laws. Constraints on the design come from human anatomy, the needs of the target customer, and the intended price.
2. Designing. This is the development of a detailed design for a garment. It usually proceeds by the designer making successive modifications to the design of a previous garment (her own or someone else's), either with sketches on paper or in her imagination. (Designers in all fields including knitwear divide into two types: visualisers who use drawings only for communication, and drawers who use sketches to represent their ideas as they develop them (Waddell, 1992).) Designers communicate their ideas with sketches, swatchesor verbal descriptions of differences from other garments. (A swatch is a piece of knitted fabric produced as an example). Designs for stitch structures are sketched or knitted by the designer by hand or with a manual machine. (A stitch structure is a combination of stitches that make up a pattern, comprising a certain number of rows and columns, which may then be repeated.) At this point a decision is made about whether to work the design out in detail and produce samples. Garments using only simple stitch structures can be worked out completely on paper without involving technicians, but garments with more complex stitch structures can only be designed in detail (sensibly) in collaboration with the technicians. Designers work on their own, even when they have colleagues, and usually under time pressure, which is sometimes intense.
3. Entering Jacquards. This is the programming of a stitch structure in the programming formalism of a particular knitting machine. This is typically the transfer of a design on paper, but simple stitch structures can programmed directly. When they are programmed they are called Jacquards. (The term originally meant a specific type of multi-colour pattern, by analogy to Jacquard looms.) Who does this depends on the company: sometimes the designer and sometimes the technician; a few companies employ someone specially to enter Jacquards. If a design for a stitch structure is selected for further development, the technician has to knit the fabric on a power machine with the intended yarn to determine the height and width of the pattern in real life, before its placing on the shape can be worked out in detail.
4. Programming. The knitting machine technician's primary job is to program the knitting machine to knit the design produced by the designer, but to do so at a reasonable cost (in other words, as fast as possible while minimising problems and producing acceptable results). This requires a detailed knowledge of the capabilities of the knitting machine and the limitations of what can be achieved with a particular yarn, as well as competence with the CAD system. The most complex programs can take up to two weeks' work to develop. Programming knitting machines requires the same type of thinking as programming computers in assembler, especially in that one has to keep track of the states of a lot of different locations in the machine, and modify values and move them around as efficiently as possible. Before the advent of electronically controlled knitting machines, the programs for flat bed machines were implemented on rolls of two metre long metal punch cards; at this time the designers had no contact at all with the technology.
5. Make Up. The final stage in the development of the garment is the production of a prototype, using the fabric produced by the program on the industrial knitting machine, which is done by the sampling make up staff (or sometimes the designer herself). This includes making cutting patterns for the pieces of a cut-and-sew garment, according to the specifications provided by the designer or the intended customer (usually a buyer for a retail chain). We omit discussion of the difficulties involved in this.
The development of a garment that is both an acceptable design and that can be produced at the right cost often involves a lot of backtracking. Almost all machine knitted garments are produced to strict price constraints, and the length of time a garment takes to knit is a major determinant of its cost. (The latest power machines can knit all sorts of fancy structures, but too slowly for them to be commercially viable.) So the technicians have to devise the most efficient programs possible, and tell designers if their designs aren't economic at particular price point. Designers often come up with stitch structures that cannot be knitted on an industrial machine, or that are too expensive to produce (that is, too slow or problematic to knit). Usually the technicians then work with the designers to reach a good compromise between cost and appearance. This can involve a lot of iteration, as can finding good ways to place patterns onto shapes. Because of the strong time constraints they work under, technicians sometimes make major modifications to the design without consulting the designer, without much concern for the appearance of the result.
The advent of knitting machines capable of producing much more complex stitch structures, combined with the development of CAD systems that designers can use (for only part of their work, and only if they get access to them) has created an overlap between the roles of designers and technicians that has never existed before.
The technicians we have talked to have invariably told us that their task would be made very much easier if the designers had a much greater understanding of power knitting machines and their CAD systems. The designers would then understood what was feasible and cheap and what was not, and could design accordingly. Experienced designers and technicians also comment that some technicians tell designers that something is impossible when it can be done, because they are too lazy to program it, and more technical knowledge would protect designers against this. We have also been told by many designers that they would like to have a much better understanding of the technology, but they are never properly trained either during their design courses or in their companies, and have very little access to CAD systems. The testimony of the technicians is compelling evidence that garments could be designed faster, if the large amount of iterative modification could be reduced by giving designers better training and more access to CAD systems, so they could develop designs they knew were feasible and cheap to a more advanced stage. If this is so, better garments could be produced if the time saved could be devoted to more careful refinement of detail and sizing. Some of this benefit should come from greater technical knowledge even without greater access to CAD systems. Later we discuss reasons why trying to improve the efficiency of the design process by giving designers more technical training is not unproblematic.
4.The Knitwear Industry
To provide a context for our discussion of the attitudes, social issues and economic factors involved in the organisation of the knitwear industry and its use of computer technology, we describe the different groups of workers in the industry. (A lot of this section is inapplicable to the very small companies owned and run by a designer; they are far too small to afford the expensive technology we are concerned with in this paper, so largely irrelevant to the issues we are discussing.)
Knitwear Designers. Almost Entirely Female. Male fashion designers are a small but significant minority, but male knitwear designers are quite rare, and we have heard of exactly one man taking a specialist knitwear design course. Almost all knitwear designers in Britain are graduates of design courses at the former polytechnics, which last three years plus a year working in industry. (The polytechnics were set up to provide more practically oriented higher education than the universities; in practice their courses are very similar to those at universities, though they tend to attract less able students; they were permitted to call themselves universities in 1992.) The designers see themselves as middle class graduate professionals doing a poorly paid professional job. (Starting salaries are usually under £10 000, which is much less than for almost all other graduate jobs.) A large majority of knitwear designers are under 30. (Those who returned our questionnaire were the most senior in their companies, and had an average age of 29; most had taken courses including a knitwear component, but few had done specialist knitwear courses.) The stereotype among other workers in the knitwear industry that "designers leave to have babies" appears to have some foundation in reality: most give up designing after a few years, though many switch to other jobs; designers returning to work after starting families usually go into management or work in other occupations. The average time a designer remains with one employer is about three years; our informants comment that designers tend to get stale designing for one market after about two years. In consequence designers are usually younger and more junior than the people they work with, and are not regarded as long term investments. This pattern is also influenced by what companies want in a designer: as one designer commented to us, 'youth is considered a positive asset in designers, as they are more "in touch" with changing ebbs of fashion and are seen as "fresher" and not yet "jaded". '