An examination of the impact of new technology on industrial designers’ search for inspirational materials S.J. Westermana, S. Kaura, C. Mougenotb, L. Sourbeb, & C. Bouchardb a. Institute of Psychological Sciences, University of Leeds, Leeds LS2 9JT UK b. Societe D’etudes et de Research de L’Ecole Nationale Superieure D’Arts et Metiers, Paris, France Abstract This paper considers the impact of new technology on creative industrial designers’ ability to locate inspirational materials. Creative design is identified as having particular information retrieval requirements. The importance of diversity and structure in search results are examined in this context. The potential for designers’ (users’) search strategies and the model of information display to provide serendipity and structure is discussed. Results from interviews with designers and tests of student designers are mentioned in support, and areas for further work proposed. Keywords: creativity, design 1. Introduction

need for communication between team members and a shared design vision (see e.g., [5]).

The search for inspirational materials is an important part of the process of creative industrial design. In this paper we explore key features of this task component and consider how it is affected by the introduction of new technologies. To support our comments, in addition to reviewing relevant literature, we draw on data gathered from interviews that we have conducted with designers (mostly car designers (n=22), but also a small number of designers from multimedia and graphic design (n=3), packaging (n=1) and furniture design (n=1)) [1]. We also refer to results from laboratory studies in which design students were asked to perform keyword searches for images as a precursor to fulfilling an imaginary design brief [2][3]. Implications are considered generally with regard to computer-based support and, in particular, the TRENDS system (see www.trendsproject.org) that is being developed to provide industrial designers with improved access to inspirational visual materials.

To produce creative solutions that are compatible with the intended context of use, designers need to be able to work within these constraints. For most designers an important part of this process involves the identification of inspirational materials that can be used as a creative ‘springboard’ (see e.g., [6]). These materials can also be useful in dealing directly with constraining factors by helping to communicate ideas to colleagues and customers [7][8] and supporting an iterative and user-centred design process. “Sources of inspiration play a number of important roles in design thinking, as definitions of context, triggers for idea generation, and as anchors for structuring designers’ mental representations of design” [7]. Interviews that we have conducted with designers support this position, with reports of image boards being used to communicate ideas to clients and colleagues, and collections of images being gathered and used as reference points for designs.

1.1 Design constraints There are a number of constraints placed on industrial designers that tend to inhibit creativity (see [4]). For example, designs tend to be created with reference to a specific context of use that is set out in the form of a design brief that details certain physical product characteristics (e.g., dimensions), anticipated users, and key brand values with which the design should be compatible. Sometimes designs must be developed as part of a team and this produces additional constraints in the form of the

In this paper we are concerned with the strategies that designers use to locate computerbased inspirational materials, and the influence that new technologies many have on this process. Digital archives are very useful for design [9], and new technology presents new opportunities for locating, storing and displaying materials. However, new technologies may also constrain creativity [5] by introducing a logic that is incompatible with the synthetic nature of the design process [10]. Again, results of interviews that we have conducted with designers support this concern. Some designers (although a relatively small number) think that the design process should be natural and uncluttered, and that new technology is not compatible with this. For these designers simplicity and the use of

‘traditional’ techniques are important. This view was perhaps more prevalent in older designers (those with greater job experience), but this was not exclusively the case. 1.2 Serendipity Definitions of creativity refer to two key determinants. Creativity requires that the product is regarded as both novel and of value (see e.g., [11]). In this sense, creativity requires a deviation from the norm. Divergence, ambiguity [10], and even risk [12] can be important contributors. “To live a creative life we must lose our fear of being wrong” (J.C. Pearce: cited by [13]) Recently, in the context of computer-based information systems, the advantages of serendipity have been discussed (e.g., [14]). On one view the association between serendipity and ‘user happiness’ can be described by a ‘serendipity curve’ with intermediate levels of serendipity being preferred, too little serendipity being boring, and too much resulting in un-interpretable chaos [15]. Results of our interviews with industrial designers indicate that, in the context of creative design, they have recognised the value of serendipity and have taken advantage of it on occasions. If serendipity is advantageous in this context the question then arises as to how it can be thoughtfully be controlled as part of a computer-based image retrieval system. A degree of divergence, between query and image retrieved, is inbuilt (although generally held to be unwanted) in the form of the ‘semantic gap’ [16]. A working assumption is that this will take the form of non-systematic error, producing random variation in search results. However, contrary to the creative designers’ requirements for serendipity in search results, the goal of ‘traditional’ information retrieval systems is to achieve perfect recall and precision [17]. In coming years, as image retrieval algorithms improve, so this variation will be reduced. Regardless of degree, identifying this source of serendipity constitutes acceptance or, at best, measurement, rather than control. Divergence in results retrieved by the system can be produced by the accidental or intentional actions of the creative designer (system user) [18]. A relevant search strategy, that can be employed when using computer-based keyword information retrieval, is to ‘under-specify’ the search by using few key words. Consistent with this, in studies conducted with design students we have

found that 50% of the sample chose not to use the maximum permitted three keywords to define their search. Another method for providing the user with a degree of control over serendipity, favoured by Beale [14], is to present the results of a keyword search in the form of a multi-dimensional spatialsemantic visualisation (see e.g., [19]) that allows the user to browse for inspirational images by navigating the space. This would also enable the user to identify additional co-varying image properties that had not occurred to them at the time of entering the search terms. In interviews, designers have been enthusiastic about the possibilities of using semantic-spatial mapping of image properties and consequently browsing of database content will be one of the features of the TRENDS system. Specific details regarding the identification of dimensions have yet to be determined. However, it is envisioned that, following a keyword image search, images will be a displayed as ‘thumbnails’ and located according to either: i) the semantic match of the image to the query; or ii) the semantic match of each image to others in the retrieved set. Whether images are displayed on two or three axes (dimensions) may be the subject of empirical testing (cf. [20]). 1.3 Structured divergence Information that diverges from the norm can be regarded as necessary but not sufficient to ensure a creative result. The designer must be able to recognise similarities between information retrieved and the problem domain (i.e., the design brief). Related to this, several current theories of creativity emphasise the need for both divergent and convergent thinking with an imbalance in either direction having potentially detrimental effects [21]. For creativity, it must be possible for the designer to draw analogies between the ‘inspirational domain’ and the ‘task domain’ (see e.g., [4]). The difficulty associated with this will be a result of the semantic distance between the two (see e.g., [22]). A significant element of creativity may result from an ability to detect similarities between apparently dissimilar materials (cf. [23]). Arguably, the extent of a designers’ creativity relates to the degree of divergence they can ‘sustain’ (i.e., the distance between sources of inspiration and the design problem) while still being able to identify commonalities between domains, although this may

also be influenced by expertise. It is to be expected that design experts will be better able to construct relevant links to the design brief when retrieved images are semantically more distant [7][22]. There are benefits for the quality of design when designers are able to form ‘rich’ patterns of cognitive association between design constructs [24]. This is consistent with associational models of creativity (e.g., [25]).

location, superimposing and fusing them”. The approach was identified as influential in the thinking of da Vinci, with Rothenberg concluding that it produces analogical thinking, but also outcomes of which the designer is unaware.

An important strategy, adopted by designers as a means of achieving structured diversity when locating inspirational materials is to search for materials in organised semantic domains that bear some relevance to the target domain. For example, interviews with car designers identified fields such as architecture, furniture, music, animals as frequent sources of inspiration [1]. In laboratory studies of image retrieval we have found that participants sometimes use one car-related keyword to ‘anchor’ their search to the design domain, but then couple this with more semantically distant search terms (e.g., car exciting futuristic; bold colourful cars). This may produce a somewhat similar effect to domain search, but for more abstract constructs.

The sample of designers that we have interviewed is heavily biased in favour of car designers, and therefore we need to be cautious when making inferences about different requirements of different design disciples. Generally, there was a good deal of consensus on major issues between designers, both within and between disciplines. However, there are also some potentially important differences in responses between design disciplines that are beginning to emerge that would influence the nature of the computer-based facilities required. For example, media/graphic designers want to be able to have ownership of images that they locate, so they can use this in work that they produce. They currently use websites such as www.istockphoto.com that provide this facility. In contrast, this is not a requirement for other types of designer.

Nature is a particularly important domain source of inspiration for many designers [1][6][7]. “All art is but imitation of nature” (Lucius Annaius Seneca: cited by [13]). Natural forms may have functional advantages, having evolved over long periods [26]. The study of ‘biomimetics’ is based on this premise: examining how designs can be developed using solutions that are present in nature. Use of analogy rather than duplication is thought to be the more successful route to take [27]. Such design solutions may also have aesthetic advantages perhaps arising from an economy of form [26]. However, [28] warns that we should be careful about being too quick to infer that designs are derived from natural solutions. Correlation is not causality. It may be that nature and designers have independently arrived at an optimal solution. An example of another strategy designed to produce structured divergence is described by Mete [6] in the context of ready-to-wear clothing. This involves designers superimposing images of two different fashion designs, one on top of the other (this is an example of a “homospatial process”). Inspiration arises from the combination of the two. [29] studied this in the context of the visual arts, concentrating on a range of artists including Leornado da Vinci. He found visual metaphors arose from “discrete entities occupying the same spatial

1.4 Requirements of different design disciplines.

For some design disciplines, fashion trends are largely determined by industry leaders, rather than the consumer. For many packaging/product designs, due to the long lead time on product materials, decisions need to be made well in advance. Moreover, particularly for smaller design agencies, it can be commercially risky not to ‘stay with the pack’ in terms of design trends. Doing so also means that materials are more readily available and at a cheaper price. For these designers inspirational materials need to reflect these predetermined trends. The integration of multi-sensory materials is more important for some design disciplines. In large part this is determined by the nature of the materials involved in the design. For example, in the context of the design of car interiors designers have told us during interviews that swatches of material are an important source of inspiration. For other industries, e.g., the fashion industry, tactile properties of materials could play an even larger role (see [6]). However, this is not an insurmountable problem for new technologies. Computer-based systems such as the TRENDS system do not need to replace all existing aspects of the design process. They simply need to be able to integrate with those that remain unchanged.

[3]

Westerman, S.J. & Kaur, S. (submitted). Creative industrial design and computerbased image retrieval: The role of aesthetics and affect.

[4]

Bonnardel, N. (2000). Towards understanding and supporting creativity in design: Analogies in a constrained environment. Knowledge-based systems, 13, 505-513.

[5]

Kalay, Y. E. (2006). The impact of information technology on design methods, products, and practices. Design Studies, 27, 357-380.

[6]

Mete, F (2006). The creative role of sources of inspiration in clothing design. Int Journal of Clothing Science and Technology 18 (4) 278 - 293

[7]

Eckert, C. & Stacey, M. (2000). Sources of inspiration: A language of design. Design Studies, 21, 523-538.

[8]

Van der Lelie, C. (2006). The value of storyboards in the product design process. Personal and Ubiquitous Computing, 10, 159-162.

[9]

Buscher, M., Friedlaender, V., Hodgson, E., Rank, S., & Shapiro, D. (2000). Designs on Objects: Imaginative practice, aesthetic categorisation, and the design of multimedia archiving support. Digital Creativity, 11(3) 161-172

[10]

Diffrient, N. (1994). Confessions of a reluctant technologist. Design Management Journal, 5, 83-85.

[11]

Eysenck, H.J. (1994). The measurement of creativity. In M.A. Boden (Ed.), Dimensions of Creativity. Cambridge, MA: MIT Press.

[12]

Richmond, W. (2000). The creative process. Communication Arts, 42, 6, pp. 230.

[13]

Sebastian, L. “Sources of Inspiration”. http://www.specialtiesgraphics.com/infobit s/ftouch/issue9/inspiration_issue9/inspiratio n.htm. Accessed 30/04/07.

[14]

Beale, R. (2007). Supporting serendipity: Using ambient intelligence to augment user exploration for data mining and web browsing. International Journal of HumanComputer Studies, 65, 421-433.

1.5 Conclusions The task of creative industrial design places particular demands on computer-based image search facilities. What is normally considered to be an effective information retrieval algorithm (i.e., producing high precision and recall, and closing the ‘semantic gap’) may be one that is rather poor for serendipity, and thereby creativity. This awaits detailed empirical evaluation and consideration of both optimal variability in search results (‘noise’) and also optimal shape of distribution (e.g., a few outlying values may be sufficient for this purpose). Regardless of this, designers do have search strategies at their disposal that can induce serendipity, e.g., using poorly specified searches. Moreover, the development of interface display formats that provide a semantic-spatial mapping of results and enables the user to browse, seem to be a popular and potentially valuable alternative means of allowing serendipitous identification of items. Computer-based image retrieval may have strong advantages for supporting certain elements of the design task. However, there are other components, e.g., providing access to tactile materials, that would be problematic. For this reason it is important that systems to support designers are developed with a clear view of the design task, so they can be successfully integrated with other components. Acknowledgements This research is being conducted as part of the TRENDS project, which is supported by the EC FP6. This paper provides an overview of work that has been reported in more detail and submitted for publication elsewhere [1][2][3]. References [1]

[2]

Kaur, S., & Westerman, S.J, Mougenot, C., Sourbe, L., & Bouchard, C. (2006) Computer-based support for creativity in industrial design. Poster presented at the CCID Conference 2006. Westerman, S.J. & Kaur, S. (submitted). Supporting creative industrial design with computer-based image retrieval.

[15]

[16]

Usernomics. “Add a little more random to your product”. http://www.usernomics.com/news/2007/01/ add-little-more-random-to-yourproduct.html. Accessed 30/04/07. Enser, P. & Sandom, C. (2003). Towards a comprehensive survey of the semantic gap in visual image retrieval. Lecture Notes in Computer Science, 2728, 291-299. Heidelberg: Springer-Verlag.

creative design: A cognitive approach. International Journal of Human-Computer Studies, 63, 422-435. [23]

Liu, Y-T. (1995). Some phenomena of seeing shapes in design. Design Studies, 16, 367-385.

[24]

Goldschmidt, G. & Smolkov, M. (2006). Variances in the impact of visual stimuli on design problem solving performance. Design Studies, 27, 549-569.

[25]

Pereira, F.C. & Cardoso, A. (2002). Conceptual blending and the quest for the holy creative process. In Proceedings of the 2nd Workshop on Creative Systems. ECAI 2002, Lyon, France. Podborschi, V. & Vaculenco, M. (2004). Natural shapes: A source of inspiration for eco-design. In D. Talaba & T. Roche, (Eds.), Product Engineering: Eco-Design, Technologies, and Green Energy. Springer.

[17]

Baeza-Yates, R. & Ribeiro-Neto, B. (1999). Modern Information Retrieval. New York: Harlow: Addison Wesley.

[18]

McLachlan, F. & Coyne, R. (2001). The accidental move: Accident and authority in design discourse. Design Studies, 22, 8799.

[26]

Westerman, S.J., Collins, J., & Cribbin, T. (2005). Browsing a document collection represented in two- and three-dimensional virtual information spaces. International Journal of Human-Computer Studies, 62, 713-736.

[27]

Vogel, S. & Vogel, J.G. (2000). Copying life’s devices. Current Science, 78, 14241430.

[28]

Vogel, S. (2000). Rhino horns and paper cups: Deceptive similarities between natural and human designs. Journal of Biosciences, 25, 191-195.

[29]

Rothenberg, A. (1980) Visual Art. Homospatial Thinking in the Creative Process. Leonardo 13 (1) 17-27

[19]

[20]

Westerman, S.J. & Cribbin, T. (2000). Mapping semantic information in virtual space: Dimensions, variance, and individual differences. International Journal of Human-Computer Studies, 53, 765-787.

[21]

Cropley, A. (2006). In praise of convergent thinking. Creativity Research Journal, 18, 391-404.

[22]

Bonnardel, N. & Marmeche, E. (2005). Toward supporting evocation processes in