machines requires the same type of thinking as programming computers in assembler, ..... such as industrial designers, notably automotive designers, make.
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, swatches or 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". ' Knitting Machine Technicians. Almost Entirely Male. We have not yet heard of a female knitting machine technician. Technicians are trained in in-house training schemes at large companies, and are usually recruited straight from school or by training up bright young knitters. They see themselves as mechanics: working class skilled manual workers promoted to good jobs. They usually remain at one company for a long time, often for their entire careers. They are usually older than the designers they work with. Sampling Make-Up Staff. Almost Entirely Female. We have heard of no males working in sampling make-up. Sampling staff are recruited by promoting the most able production make-up workers. They see themselves as working class skilled manual workers promoted to good jobs. Knitting Machine Operators (Knitters). Almost Entirely Male. We have heard of female knitters but they are uncommon in Britain. Knitters are responsible for running groups of industrial knitting machines. Production Make-Up Staff. Almost Entirely Female. The proportion of male make-up employees depends on the company: we know of one with several men, but all the others we have information on have none. The make-up employees comprise garment cutters, who cut knitted fabric into pieces, and people who operate various kinds of sewing machine, principally overlockers. Packing and quality control is also usually done by women.
Managers. Predominantly Male. Most managers are recruited from people with no background in the knitwear industry, or who have textile management or textile technology degrees, though some are promoted from within the company. According to our informants very few managers are former designers or technicians. This has the consequence that companies are managed by people lacking a detailed first hand understanding of the design process. Knitwear is dominated by small firms (Rubery et al., 1992). A large fraction of companies are family concerns run by their owners, where the managerial structures depend on the family structures; most of the other larger mass-market companies are subsidiaries of large organisations like Courtaulds and Coats Viyella (which are the two giant companies with many semi-independent branches). Only the larger knitwear companies have enough designers to require a head designer with managerial duties. Recently two very large companies appointed their head designers directors: this was reported to us by various people as a complete novelty. One designer-turned-manager commented to us that very few designers have the ability and personality to be managers, and that her company wanted to recruit a designer with management potential and had difficulty finding one. Many writers have commented that women are discriminated against in promotion decisions (for example, Alban-Metcalfe and West, 1991; Bruce and Lewis, 1990); knitwear designers have no competition from men for promotion as designers, but we expect that sex discrimination affects their prospects of promotion to more general management jobs. However we are unable to unpack the influences exerted by sex discrimination, discrimination by group stereotype, the intellectual abilities of designers and other potential managers, designers' personalities, and the differences of priorities and values between designers and managers. The knitwear industry would be a good domain for a serious study of the interaction of these factors. Neither designers nor technicians have much power within knitwear companies, nor influence over their decision making processes, but technicians gain some influence by being difficult to replace. The designers' major source of influence is their special knowledge of future developments in fashion and the requirements of the market, though they often have close ties to customers and suppliers. They are often seen as both junior and temporary, and very easy to replace (as there are more young designers than jobs). This makes managers reluctant to invest money, time or effort in training designers, especially in how to use their companies' CAD systems. (Almost all the designers we talked to expressed an interest in more CAD training than they were able to get, even in the biggest and poshest companies where they have most respect and freedom.) Decisions about which knitting machines to buy are made primarily on price and functionality. However companies are very reluctant to change machine manufacturers unless their current manufacturer cannot provide the functionality and service they need, partly because the expertise of their technicians is to some extent limited to one manufacturer. The designers' only influence on CAD system purchasing decisions comes from knowing what functionality the coming fashions will require. This has the potentially important consequence that there will seldom be much pressure to choose more designer-friendly CAD systems. We have not investigated the causes of the extremely sharp sex divisions in the knitwear industry, as our brief has been understanding designers, but we should make the following comments. The reasons why knitwear designers are almost exclusively female are rather different from the reasons why the other occupations in the industry have such sharp segregation. Male knitwear designers do not have to face strong social barriers to finding jobs, but men wanting to be knitwear designers appear to be rare. The occupations of knitter and make-up worker (seamstress) grew out of traditional crafts with sharp sex divisions in the
industrial revolution; the persistence of such division in trades originating in crafts has been well documented in other industries, and attributed partly to a policy of recruiting to fit into existing social groups (Aitkenhead and Liff, 1991). Knitting machine technician is an occupation requiring mechanical aptitude, and before the days of electronic machines and computerisation it involved a significant amount of work requiring physical strength (especially moving rolls of metal punch cards, which were heavy). The use of the belief that women are incapable of hard physical work, to exclude women from technical occupations, has been documented recently in other industries by Cockburn (1985); but the hard physical work has been eliminated from the work of knitting machine technicians, which is largely programming and tweaking the controls on knitting machines. Cockburn argues that women are actively excluded from technical fields because men working in them regard technical competence as a masculine attribute fundamental to their self-identities as males. We have little direct evidence on which to base comments on the applicability of this view to the knitwear industry, and none on companies' recruitment policies for technicians. But we conjecture on the basis of our conversations with technicians that a more important factor influencing the sex division is the unconscious bias produced by managers' male images of typical or successful technicians. Webb (1991) remarks that 'It is practically not possible to make an absolute distinction between the job and the job occupant: job requirements are defined in terms of current job holders, including gender.' When asked the question "why aren't there any female technicians?", the technicians we talked to replied "I don't know" or "there aren't any", but did not express hostility to the idea. One experienced technician with whom we discussed this issue commented to us that women used to be regarded by people in the industry as incapable of work requiring mechanical aptitude or strength. He had no objection to the idea of women becoming technicians, and knew one female textile technology student who wanted to, but he thought she would not get a chance, not because of such raw explicit sexism but rather because potential employers would regard her as too much of a gamble, because she did not fit their image of a technician, and would be impossible to dismiss if unsuccessful.
5.
The State of CAD System Technology for Knitwear
At the moment commercial knitwear designers make almost no use of computers in the earlier stages of design. The extent to which computers are used for design varies widely between design domains, and this is not the place for a review, but we should note some significant parallels. One is that designers concerned with the design of complex three dimensional shapes subject to various constraints, such as industrial designers, notably automotive designers, make very little use of CAD systems (Tovey, 1989, 1992). Automotive designers are male and in a high prestige industry comprising huge rich companies, so to some extent they constitute a control for the social and economic factors we describe in this paper (though the CAD systems used in car design are so expensive that even Ford think twice before buying them). The exception to this pattern is architecture, where shape design is done using CAD systems; but architecture is largely about the representation and manipulation of simple geometrical shapes, which is relatively easy for a computer to do, and a huge amount of research has been devoted to computerising it (for example Bijl, 1989; Liu, 1991). Perhaps a more significant parallel to industrial design is that both industrial designers and knitwear designers hand over their projects for others to do the detailed technical specification; CAD systems are used extensively by designers in fields where they carry projects from initial conception to detailed technical
development, such as architecture, textile design (Devane, 1992) and typographic design (Hewson, 1994). The CAD systems in use in the knitwear industry are developed by manufacturers of industrial knitting machines for use with their own machines. (In addition to the machine builders, Minima, a software house based in Reutlingen, produces a CAD system for programming knitting machines, which is compatible with the machines produced by all the flat bed machine manufacturers: Universal, Stoll, Shima Seiki, etc.) These CAD systems are designed to increase the ease and efficiency of programming, and not at all to support the invention of the designs themselves. They are primarily visual programming environments, and are intended for the technicians to translate completed designs into runnable programs. Why have the producers of CAD systems for knitwear neglected the designers? Both the objective characteristics of the industry and the available technology, and the attitudes of the system developers play a part, but we argue that the former are much more significant. Nevertheless the technology has progressed to the point where supporting the designers has become a relevant issue, and the development of systems that suit knitwear designers is held back by the attitudes and beliefs about knitwear designers and CAD systems held by system developers, knitwear company managers, design teachers and the designers themselves. Knitting machines need to be programmed somehow, so the machine manufacturers must provide some way to do it, and the quality of the support provided for the programming is a major selling point for the machines. This is the driving force behind the development of better CAD systems for knitwear. The knitting machine companies have constant contact with technicians using their machines, and can see how well the technicians can learn to use their systems in the training courses they provide. Conversely, having technicians who can program knitting machines well and efficiently is a clear economic necessity for knitwear companies, who will usually pay for them to attend training courses for new systems. This aspect of the business has not changed since the days of metal punch cards, when the division of responsibility between designers and technicians was absolute. The state of CAD technology also influences the emphasis given to supporting technicians rather than designers. Computer control of complex machines is a well worked out area of engineering competence, and producing good programming environments is very much easier than producing good CAD systems for informal design work. The first reason is that the discrete nature of knitting patterns (which are matrices of smallish numbers of stitches) makes it relatively easy to represent formally stitch structures and algorithms for producing them in a variety of different ways. Universal and Shima Seiki use colour coding; Stoll uses symbols to represent individual stitches as part of a language derived from BASIC, and has introduced an iconic representation of predefined stitch structures. The second reason is that knitting machine programming is a job that fits comfortably into the task-order constraints imposed by relatively simple software systems, whereas design doesn't. A third is that unsolved hard problems are involved in providing good CAD support for sketching and geometric design that permits imprecision and progressive refinement of details, and avoids premature commitment and hidden consequences of decisions (Tovey, 1992, Scrivener, 1993). The analogy to other design tasks where similar patterns of computer use obtain supports the view that the relative difficulty
of supporting designers is largely sufficient to cause the existing imbalance in computer use (Tovey, 1989, 1992). One almost entirely male group of computer programmers and mechanical engineers develops systems for another almost entirely male group of technicians, but we think the sex division is less significant than a personality type division. The system developers understand the thought processes and priorities of the technicians because both are programmers, but have difficulty understanding the attitudes, values and working methods of the designers. One illustration of this is the colour coding of stitch types used in the Universal and Shima Seiki CAD systems: they are effective formalisms for visual programming, but (as the first author has observed herself) they are unpleasant for people trained to be exceptionally sensitive to colour combinations. Computer scientists need detailed study to understand designers, as we know from our own experience (Eckert and Murray, 1993), as well as that of others (Scrivener,1993) and can go seriously wrong. Ashby (1992) observes that some CAD systems for textile design try to force designers into using alien working methods; while Waddell (1992) comments that some fashion design CAD systems are based on a misapprehension of the nature of the design process, often coming from a failure to understand what the designers' sketches are actually for, with the result that fashion designers find them useless and feel themselves excluded from the technological age. Nevertheless the knitting machine programming systems are being developed into true CAD systems with features that make them useful for supporting design. All three manufacturers are working on automated programming techniques and increasingly sophisticated ways to model and display the output of the knitting machine, which already include three dimensional representations of what a knitted stitch structure will look like. The development of these display facilities creates the potential for a redivision of the design task between the designer and the technician, a process that has begun at the companies where the designers enter Jacquards. (While knitwear design might benefit from the techniques being developed for computer supported cooperative working (CSCW), this would require a further shift in the conceptualisation of the design process, from one in which the technician takes over from the designer, to one where they work in parallel. We think this is still several years away.) Despite this rapid technological progress, the developers of knitting machine CAD systems have had very little interest in supporting designers until very recently: system developers at all the three major companies we have talked to told us that their companies regarded supporting designers as of minor importance or not their business. One who thought his company should pay some attention to designers said his was a minority view. One objective ground for this prevailing attitude is the absence of existing demand for designer-friendly CAD systems. In February 1994 Shima Seiki introduced a new version of their CAD system with improved automatic programming facilities and a cutting pattern construction system, which they are marketing as targeted at designers, though it is a linear development of their previous systems and designed primarily for technicians. We found it still very directive in imposing a style of working on the user. Stoll have also started marketing their most advanced systems as having designers as their main user
6.
CAD Systems in Industry
Economic factors severely limit designers' access to the existing CAD systems in knitwear companies, and so indirectly limit the demand for designer-friendly systems. The most important factor influencing the use of CAD systems for knitwear, apart from the state of the technology, is simply that the CAD systems are very expensive. For example, the Stoll systems start at £16 000, though additional systems cost £13 000 for the hardware, with free copies of the software, and go up; employing a newly graduated designer for a year is cheaper. The cheapest Shima Seiki CAD system costs £32 000 (which is the cost of a small house in Leicester, where much of the British knitting industry is based), with no site licences or options to buy the hardware separately. The most elaborate state-of-the-art systems, which are the ones with the features most useful to designers, cost up to £134 400. Neither companies nor colleges can afford to provide up-to-date systems for tasks that are not plainly essential; and only the colleges with specialist courses have knitting machines with CAD systems at all. Using CAD systems to design garments, rather than just to program knitting machines, can only be economically viable if the designers' increase in productivity pays for the depreciation costs of the machines they use; this will remain a vital consideration for the forseeable future. For knitwear companies, investment in giving designers training and access to CAD systems is not clearly essential, even if it might bring benefits, so companies short of money (which most are) will not pay for it. Thus designers often have to content themselves with the technicians' hand-me-downs, which are inevitably less user-friendly and less well-suited to supporting design; they only get access to up-to-date systems when they aren't needed for anything essential. Designers in industry report that they find it very hard to get themselves sent on training courses; this is influenced both by the lesser urgency of giving training to designers compared to technicians, and by the expectation that designers will not remain with the company long enough to return investment in training them; as we have noted earlier, designers work much of the year under enormous time pressure, making it impossible to release them for courses. There may also be sexual discrimination: Hammond (1986) and Colwill and Josephson (1983) report that women find it harder than men in the same occupations to obtain company sponsored training; however we have no easy way to separate this from other factors. However we should add to this pessimistic picture the point that companies' attitudes vary widely, and the knitwear industry as a whole is a moving target; some companies are training their designers on CAD systems and experimenting with computerised design. As a result of these economic factors, it is very difficult for most designers to make as much use of the CAD systems as they could, which limits the redivision of work the CAD systems are beginning to make possible, and limits the commercial demand for designer-friendly systems. Moreover, the difficulty designers have in gaining training and practice with CAD systems serves to prevent them from gaining the benefits to be had from a greater understanding of the potential and the limitations of their companies' knitting machines. The limited provision of on-the-job training places the onus for adequate education in computer aided design on the designers' design courses. The knitwear design courses we have studied at De Montfort University do not get students beyond basic keyboard skills and how to enter Jacquards in one old CAD system, in contrast to some textile design courses, which have embraced computer aided design (Devane, 1992). Although the computer education at De
Montfort University is not as good or as deep as we think it should be, the major limiting factor is cost: compare the annual equipment budget for the entire Textiles Department, which is £30000, to the prices for CAD systems we quoted above. Colleges without specialist knitwear courses cannot afford up-to-date knitting machines at all, so their students emerge with no experience at all of modern technology.
7.
Attitudes to Designers using CAD Systems
These economic and technological factors are the major determinants of computer use in the knitwear industry, and are sufficient to cause the existing situation. But another factor might contribute to the explanation of why computer technology is pervasive in one part of the knitwear design process, done by male technicians, and is almost absent from the creative work done by female designers, in contrast to textile design, where CAD systems used by designers are rapidly becoming important (Devane, 1992; Waddell, 1992). We have frequently met the belief that knitwear designers need not, indeed should not, know anything about the technology of knitwear production (a belief which is rejected by both designers and technicians). While this view is not a major cause of the existing division of computer use, it may serve to restrict any change to that division. Our survey of the knitwear industry is not extensive enough to permit an assessment of how pervasive it is, but we have encountered it often enough to conclude that it is fairly widespread and is held by some people who have significant influence. This view has been expressed explicitly by the Head of School responsible for one knitwear design degree, and it is implicit in the decisions made by the managers of many knitwear companies. One rationale for the view that designers do not need technical knowledge (sometimes given by managers to justify not sending designers on training courses) is that the design process has always worked perfectly well without designers knowing anything technological (or only what they pick up as they go along). The rationale for the stronger view that technical knowledge can be harmful, is that designing for a particular machine limits the designers' creativity, and leads them to produce stereotyped and repetitive designs. This view seems to have some foundation: we have been told by both an experienced knitwear designer and a design lecturer that thinking too much about the limitations of machines leads designers to develop habits and standard procedures that accelerate the design process but restrict its output, while they stop generating enough new ideas. (Designing for a particular market is even more limiting, which is one reason why designers change jobs relatively frequently.) And Devane (1992) reports that while her textile design students thrive on CAD systems, they become reluctant to attempt anything that is hard to obtain using the available systems. However we suspect an element of prejudice: the view that "girls aren't technically minded" was certainly prevalent in the knitwear industry before the days of computerisation. The knitwear designers themselves will tend to reinforce this, in that they are a group strongly selected by personality type, and conform to the negative stereotype that "women are intuitive rather than rational" much more strongly than the general female population, though most are quite capable of mastering the CAD systems and understanding the potentials and limitations of knitting machines. But this may be more a prejudice against a personality type than a sex: the belief that people can't be both "technical" and "creative" is very widespread. These views spring from the common but fallacious view that "creativity" is only required for artistic activities and is the opposite of problem solving, so that "being creative" cannot and should not involve problem
solving. Psychological studies of creativity (Weisberg, 1988; Greeno, 1980; Perkins, 1981; Weisberg, 1986) and of "convergent" and "divergent" thinking (Hudson, 1966, 1968) show that problem solving ability and fluency of idea generation are orthogonal abilities, and that success in most creative activities requires both creativity and the problem solving ability needed to sift good ideas from bad ones and think them through. A more extreme version of the view that systematic problem solving and creativity are opposites was expressed by a software developer at Minima, who asserted that being trained in logical thinking destroys ones ability to design; he told the first author that her background in mathematics and computer science made her incapable of being a designer. 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.
9.
Conclusions
The different jobs and roles in the knitwear industry are segregated by sex to a remarkable degree; moreover the complex technology used in the industry, including CAD systems, is controlled entirely by an almost exclusively male group, the knitting machine technicians, while an almost exclusively female group, the knitwear designers, is excluded from mastery and control of the technology. This is in striking contrast to other design domains, such as architecture and textile design, where CAD systems supporting designers already play a major role. In this paper we have tried to explain this, and tried to explore the question of what would be involved in giving knitwear designers mastery over the CAD systems. While we do not think prejudice against women doing technical jobs has died out in the knitwear industry, our observations lead us to believe that the sharp division of males and females by role is an orthogonal issue to the pattern of computer usage in the industry. While we have not studied the issue extensively, the evidence we have is that knitting machine technician is a male trade for the same social reasons that mechanical jobs in other industries have been
male, while knitwear designers are female because of the distribution of interest in an activity perceived as very feminine, and the social pressures influencing boys and girls in their choice of career. We conclude that the major determinant of the pattern of computer use in the knitwear industry is the fundamental importance (and difficulty) of programming knitting machines, which has always been the province of the knitting machine technicians. The knitting machine manufacturers need to provide ways to program their machines, and the quality of the programming environments they provide is a critical area of competition between them. Supporting designers has simply never been essential. An important secondary determinant has been the availability of relatively well worked out technology for computer controlled manufacture and for programming environments; providing good computer support for designers is a harder problem that is far from fully solved. The economics of the knitwear industry is another significant secondary determinant. The existing CAD systems are very expensive, so that knitwear companies cannot afford to buy more than the minimum number they can get away with. At the same time few companies can afford to invest resources in training the designers on CAD systems when they cannot see a clear benefit and expect the designer to leave after only a few years. We have found a widespread belief (not shared by designers in industry or by the technicians they work with) that designers do not need any understanding of the CAD systems or knitting machines, and are likely to be harmed by it. We conclude that while this does not cause the sharp division in CAD system usage, it contributes to the lack of interest on the part of machine builders in producing designer-friendly CAD systems for the future, and the lack of support and encouragement given to designers to master the existing CAD systems. The view that familiarity with a CAD system can actively harm a designer has some element of truth, but we conclude that it is outweighed by the advantages of being able to use CAD systems, and hypothesise that the negative effects can be combatted by education and by gaining full mastery of the system. The fallacious view that systematic problem solving is opposed and hostile to creativity contributes to these attitudes; we conjecture that similar views might appear and exert a harmful influence in other industries, if the development of CAD technology leads to an integration of the activities of designers and engineers. We conclude that the strong prima facie case that the efficiency of the knitwear design process could be improved by giving knitwear designers access and competence with CAD systems for knitwear, resulting in cheaper or better garments, is in fact supported by the evidence available to us, though it remains untested. But achieving this improvement would require CAD systems cheap enough for much more extensive use, and designers who emerge from their degree courses already well trained in how to use them. This in turn would require much greater emphasis on CAD systems in knitwear design courses. Further improvement in the position of knitwear designers and in the efficiency of the design process will depend on the development of CAD systems better suited to the needs of knitwear designers. Efforts are being made to develop such systems, not least by ourselves, but whether they will be used to their full potential will depend on how affordable they are, and on how well design students are taught to use them.
Acknowledgements This research was partly funded by the ACME Directorate of the SERC, grant number GR/J40331. Prof Ernest Edmonds, Dr Thomas Green, Monica Jandrisits, Wendy Nicholson and Dr Alison Green made helpful comments on earlier drafts of this paper. Our research also benefited from conversations with Bryan Murray. We are very grateful to all our informants for the time and effort they devoted to talking to us, especially Monica Jandrisits and Annabelle Duncan.
References AITKENHEAD, M. and LIFF, S. (1991) 'The effectiveness of equal oportunity policies' in J. FIRTH-COZENS and M.A. WEST (Eds) Women at work, Milton Keynes, Open University Press. ALBAN-METCALF, B. and WEST, M. (1991) 'Woman managers' in J. FIRTH-COZENS and M.A. WEST (Eds) Women at work, Milton Keynes, Open University Press. ASHBY, C. (1992) 'Design by Software' in W. ALDRICH (Ed) CAD in clothing and textiles: a collection of expert views, Oxford, BSP Professional. BIJL, A. (1989) Computer Discipline and Design Practice, Edinburgh, Edinburgh University Press. BURCE, M. and LEWIS, J. (1990) 'Women designers - is there a gender trap?', Design Studies, 11, pp. 114-120. CASSELL, C. (1991) 'A woman's place is at the word processor: technology and change in the office' in J. FIRTH-COZENS and M.A. WEST (Eds) Women at work, Milton Keynes, Open University Press. COCKBURN, C. (1985) Machinery of Dominance, London, Pluto Press. COLWILL, G. and JOSEPHSON, W.L. (1983) 'Attitudes towards equal opportunity in employment: the case of one government department', Business Quarterly, 48, pp. 87-93. DEVANE, J. (1992) 'Training and Eduction in CAD clothing/ textile systems', in W. ALDRICH (Ed) CAD in clothing and textiles: a collection of expert views, Oxford, BSP Professional. ECKERT, C. and MURRAY, B.M. (1993) CAD/CAM for Knitwear Design and Development: User Study, ACME Grant GR/J40331 D1-Report. Loughborough, LUTCHI. GREENO, J. G. (1980) 'Trends in the theory of knowledge for problem solving' in D.T. TUMA and R. REIF (Eds), Problem solving and education: Issues in teaching and learning, Hillsdale, NJ, Lawrence Erlbaum.
HAMMOND, V. (1986) 'Management training for women', Journal of European Industrial Training, 10 , pp. 15 - 22. HEWSON, R. (1994) Making and marking: how typographic designers use sketches. Ph.D. Thesis, Institute of Educational Technology, The Open University. HUDSON, L. (1966) Contrary Imaginations, London, Methuen. HUDSON, L. (1968) Frames of Mind, London, Methuen. LIU, Y. (1991) 'Schematic-Designer: a knowledge-based CAD system for schematic design in architecture', Design Studies, 12, pp. 151-167. LOCKWOOD, E. and KNOWLES, W. (1984) 'Women at Work in Great Britain', in M.J. DAVIDSON and C.L. COOPER (Eds) Working Women, Chichester, John Wiley. PERKINS, D. (1981) 'The mind's best work', Cambridge, MA, Harvard University Press. RUBERY J., HUMPHRIES J. and HORRELL S. (1992) 'Women's Employment in Textiles and Clothing' in R.M. LINDLEY (Ed) Women's Employment: Britain in the Single European Market, London, HMSO. SCRIVENER, S.A.R., CLARK, S.M., CLARKE, A., CONNOLLY, J., GARNER, S., PALMEN, H., SMYTH, M. and SCHAPPO, A. (1993)'Real-time Communication Between Dispersed Work Groups via Speech and Drawing', Wirtshaftsinformatik, 35, pp. 149-156. TOVEY, M. (1989) 'Drawing and CAD in industrial design', Design Studies, 10, pp. 24-39. TOVEY, M. (1992) 'Intuitive and objective processes in automotive design', Design Studies, 13, pp. 23-41. WADDELL, G. (1992) 'CAD for clothing and textiles' in W. ALDRICH (Ed) CAD in clothing and textiles: a collection of expert views, Oxford, BSP Professional. WEBB, J. (1991) 'The gender relation of assessment' in J. FIRTH-COZENS and M.A. WEST (Eds) Women at work, Milton Keynes, Open University Press. WEISBERG, R. W.(1986) Creativity, Genius and other myths, New York, NY, Freeman. WEISBERG, R. W.(1988) Problem solving and creativity, In R.J. STERNBERG (Ed) The Nature of Creativity, Cambridge, Cambridge University Press.
