Forum User-Centered Considered Harmful1 (with apologies to Edsger Dijkstra, Niklaus Wirth, Andy Dearden and ) [email protected] Reader in e-SocialAction Studies of IT for development have often identiªed the importance of the Communication and Research Centre of IT systems and the need for IT systems to be matched to the Shefªeld Hallam University needs of host communities. These two issues are central concerns for the Furnival Building discipline of –Computer Interaction (HCI), or Interaction Design. Shefªeld, S1 2NU Within HCI and Interaction Design, user-centered design is just one partic- UK ular view on how design processes can be organized to achieve such +44 114 225 2916 aims. This paper reports on discussions arising from a workshop held at the Computer/Human Interaction (CHI) 2007 conference in San Jose, Califor- nia. CHI is the largest HCI conference in the world. The workshop brought together a group of 45 interaction and development practitio- ners from around the world and included participants from 17 countries, including many researchers and practitioners based in emerging econo- mies such as India, China, South Africa, Namibia, and Benin. The aim of the workshop was to examine how interaction design could contribute to the success of IT for development. Although many issues were discussed, this is a necessarily selective report focusing on some of the principal themes of the workshop.

The Current Situation IT for Development is a growing ªeld of study. Ofªce-based IT systems clearly have an important role in the coordination and management of large private and public organizations and are also recognized for their use in nongovernmental, or third sector, organizations. In the domain of systems designed for commercial organizations in the private and public sector, some emerging economies, notably India, have developed their own software industries that compete for market share in the developed world. More recently, there has also been an increasing interest in how access to information and communication technologies, such as connecting to the Internet, might impact social and economic development by, for ex- ample, enabling farmers to discover ways to improve their agricultural

1. The workshop was initiated as part of the U.K. and Physical Sciences Research Council projects, “Bridging the Global Digital Divide” and “Rural e-Services.” The U.S. National Science Foundation and the Association of Computing Machinery Special Interest Group on Computer Human Interaction (ACM SIGCHI) also sponsored the workshop by providing funds for participants from developing countries to travel to attend the workshop. The work- shop was organized by Andy Dearden, Michael L. Best, Susan Dray, Ann Light, John Thomas, Celeste Buckhalter, Daniel Greenblatt, Shanks Krishnan, and Nithya Sambasivan. I would like to thank all of these contributors for their helpful in- put in creating this report.

© The MIT Press 2008. Published under Creative Commons Attribution NonCommercial-NoDerivativeWorks Unported 3.0 license. All rights not granted thereunder to the public are reserved to the publisher and may not be exercised without its express written permission. Volume 4, Number 3, Spring/Summer 2008, 7–12 7 USER-CENTERED DESIGN CONSIDERED HARMFUL output, or to obtain better prices; supporting educa- amine how the discipline could contribute its knowl- tion programs (Arora & Joseph, 2007; Dickelman & edge and experience to global development and Arias, 2007); or enabling migrant workers to com- explore the challenges that might arise in that municate with their families and friends that they effort. have left behind (Liu & Meng, 2007). Another way ICT can make an impact on poverty is by supporting Clarifying Our Objectives key activities of development organizations in areas There are questions about the objectives of design. such as microªnance (Gupta, 2007) or the develop- In HCI, a common point of reference is the concept ment of novel applications that address the eco- of usability, which the International Standards Orga- nomic, social, and personal needs of communities nization deªnes as “the extent to which a product (e.g., the ChikanCAD software used for computer- can be used by speciªed users to achieve speciªed aided design by craft weavers in Lucknow [Sharma, goals with effectiveness, efªciency and satisfaction 2007]). in a speciªed context of use.” The mobile phone has appeared as an important These criteria of efªciency and effectiveness in re- communication platform that could make a lation to speciªed goals are typically interpreted in signiªcant contribution to poverty reduction. Many terms of economic efªciency—a discourse around participants reported on the high levels of mobile individual productivity and satisfaction that often phone use in the settings where they were working risks a narrow consumerist interpretation. The work- (Marsden, 2007; Kolko & Rose, 2007; Ouko, 2007). shop participants questioned whether user-centered Kavanaugh, Reese, & Pérez-Quiones, (2007) argued design was the right metaphor for design for devel- that the mobile phone might act as a “scaffolding” opment (Bednarik et al., 2007), and concluded that technology to aid users in learning how to operate a more community-centric view was required. other computing systems. Other technologies such Introducing IT systems into a community or orga- as personal digital assistants (PDAs), digital cameras, nization is likely to disrupt social relations, bringing and programmable mobile phones have also been beneªts to some and disadvantages to others. Such used to support specially designed applications disruption is a commonplace observation in informa- (Parikh & Brunskill, 2007; Parikh & Schwartzman, tion systems literature. As Athavankar (2007) ex- 2007; Kam & Ramachandran, 2007) that contribute plains, designing for international development to community and national development. In all such cannot successfully be realized unless interaction projects, two critical success factors are the degree designers develop a rich understanding of develop- to which the technology can be easily used and ap- ment. As a result, the workshop reported back propriated by users and the degree to which the ser- to the wider conference using the slogan “User- vices offered by the technology address the primary Centered Design Considered Harmful” and sug- needs of intended beneªciaries. gested in its place a new approach of community- The discipline of HCI, or Interaction Design, is centered design. speciªcally concerned with the ways in which hu- Exploring this communal perspective is useful in mans work with and through ICT systems and how thinking about systems that are developed to meet ICT systems can be designed to ensure effective and speciªc needs of communities located in the devel- satisfying interactions. Many HCI researchers and oping world, but may also be useful in thinking practitioners (Frohlich, 2007; Parikh & Brunskill, about how a global community can engage with 2007; Gupta, 2007) have engaged with IT for devel- development. Many contributors to the workshop opment projects to design novel solutions tailored to reported on novel uses of technology to support col- speciªc local needs. In the United Kingdom, a na- laboration, discussion, interaction, mutual learning, tional research network called “Bridging the Global and support among communities spread among de- 2 Digital Divide” has been formed to create design veloped and developing regions (Diehl, 2007; Disse, responses to IT for development issues. However, 2007; Aykin, 2007; Buckingham-Shum, 2007; the workshop in San Jose was the ªrst time the Boettiger, 2007). community of HCI designers came together to ex-

2. See www.bgdd.org for more information.

8 Information Technologies and International Development DEARDEN

The Design Challenges One topic for discussion was how culture plays out in the form of the system itself and some of the International development projects raise several de- surprises that occur when technology is introduced sign challenges, and interaction design is no excep- into new contexts. For example, in Bombay, auto- tion. Most computer systems are designed for users matic teller machines for banks, where users make who are assumed (typically) to be literate and who deposits, have proven popular with prostitutes who are expected to have had some experience with use them to deposit their earnings during the night other computer platforms (e.g., keyboards, mice, so they do not have to deal face-to-face with the pointers, screens). Often assume that the bank. Ways of working need to be reºected in tech- computer will have a reliable connection to the nology design. For example, the workshop discussed Internet (i.e., electrical power and telephone con- how the importance of word-of-mouth and personal nectivity will not be a problem). Designers and users networking in ªnding work were reºected in the de- may share important elements of cultural back- sign of a recruitment portal for cleaning work in ground such as language, education, history, and India. working arrangements. These assumptions are rarely The workshop discussion identiªed the following valid when designing systems for use in the devel- three distinct processes whereby technology is oping world (Wilson, 2007; Yeo, Songan, & Hamid, shaped to support development: 2007). Toyama (2007) highlighted the danger of a variety of assumptions that technology designers • Internationalizing existing products that were (and development practitioners) might make about developed initially for a Northern market (e.g., community and individual attitudes toward new a local version of Microsoft Word) technology. • Extending indigenous social shaping of ICT- Designing for development requires systems that using existing products, but developing new are usable by people with limited literacy, who have practices around those products (e.g., using little prior experience with computers, who may not computer-aided design to support craft weav- be familiar with core concepts applied in computing ers [Sharma, 2007]) such as hierarchical classiªcation (Marsden, 2007), • “Contextual innovation” in a practice area— and whose primary language may not be the ap- designing new solutions and new platforms/ proved “ofªcial language” of the country where technologies/software in the speciªc context of they live. Furthermore, it requires systems that can development (e.g., Parikh & Brunskill, 2007) cope with intermittent power supplies and limited infrastructure, make sense in cultural contexts that and Techniques are likely to be very unfamiliar to the technology de- Although designs have proven successful in a few signers (Kotze, 2007), and overcome cultural norms contexts, most existing projects are unsustainable such as established working and governance prac- (Hettiarachchi, 2007). This phenomenon suggests tices that constrain the implementation of designs that the value delivered for the community of conceived by other cultures (Ho, Luk, & Aoki, 2007). beneªciaries is not sufªciently large to cover the Additionally, design must still consider the wider is- costs of developing and deploying the technology sue of making systems accessible to users with dis- (Kitner, Beckwith, & Ilahiane, 2007). For a , abilities (Wong, 2007; Kotze, 2007) and must be this raises the question of how to improve the de- maintainable in ªeld settings where even basic tools sign of both project and technology so sustainable such as spanners (wrenches) are difªcult to obtain development can be achieved. (Godjo, 2007; Dhewa, 2007). Designing applications One problem is that the techniques and methods that will run on mobile phones and PDAs becomes that are used successfully for designing ICT in devel- even more complex because all interaction (e.g., the oped world contexts rely upon certain assumptions display of outputs and user feedback) must occur about how users and developers can interact and through a very small interface, which in turn restricts discuss ideas that are not necessarily valid in devel- the options users have to provide input to the sys- oping world settings (Winschiers, 2007; Chetty & tem. Despite these challenges, some systems have Grinter, 2007; Medhi, 2007). One particular tradi- been designed that have successfully delivered de- tion in interaction design is the velopment beneªts.

Volume 4, Number 3, Spring/Summer 2008 9 USER-CENTERED DESIGN CONSIDERED HARMFUL school, which has concentrated on questions of may or may not be converted into practical action in how end users and other stakeholders can be more the longer term. Adopting technological solutions actively involved in the creation of software. The carries both beneªts and risks, and very vulnerable participatory design tradition is most closely associ- communities may not be ready to afford those risks ated with and development in (Kitner et al., 2007). In addition to resolving issues Scandinavian countries, where the discourse on de- of trust, a local partner can act as a cultural inter- sign has addressed questions about the power rela- preter and may help a team to avoid mistakes. tions between software developers, customers, end Design for development requires that system de- users, and other stakeholders in the software devel- signers examine their conception of development. In opment process. However, Scandinavia is recognized many cases, system designers may be unaware of as a region where social relations are exceptionally the implicit assumptions that will have a signiªcant egalitarian compared with more distanced, hierarchi- impact on the potential effectiveness of ICT for de- cal relations that are typical in many (though not all) velopment (Hettiarachchi, 2007; Athavankar, 2007). developing countries. Similarly, many techniques in In the context of researching interaction design for interaction design rely upon the words that infor- development, the software developers may have mants utter; for example, it is common practice to their own motivations and incentives (e.g., to gener- transcribe interview notes. However, in some cul- ate publications, to attract future funding) that must tures, the need to avoid loss of face can mean that be honestly acknowledged if a truly open design de- verbal utterances should not be interpreted without bate is to be possible. Thus the procedures for es- careful attention to details of the associated nonver- tablishing projects that are typical of software bal signals. creation for commerce in the developed world need It is clear that new, context-speciªc techniques to be reconsidered, as do the ethical issues that arise may be required to achieve successful design. in forming relationships with communities (Parikh & Schwartzman, 2007; Spittles & Dunckley, 2007; Bridging the Gaps Between Users Marsden, 2007). and Designers Another group at the workshop sought to iden- tify possible heuristics to guide design processes for There was general agreement that participatory ap- development and research on ICT design for devel- proaches to design are highly relevant to design for opment. It was recognized that perhaps heuristics development, but the notion of participation re- might need to be speciªc to countries or regions. quires further examination (Maunder, 2007). Following are examples of the heuristics suggested: Winschiers (2007) and Yuan (2007) argued that methods for engaging with users need to be trans- • The importance of identifying a human access lated and adapted to different cultures, just as soft- point or a local partner who is already working ware does. One workshop group focused on how to directly with the community. This may be achieve “culturally appropriate community engage- through family relations or through other ment,” dealing with issues such as intra-community mechanisms such as NGOs. diversity, power relationships, and expectations. • Doing groundwork and homework before en- They reported that in their experience it is often tering the ªeld—you should know a fair necessary to have the support of local leadership or amount about the area, and the work of the local government ofªcials before trying to work di- community before you contact them. This rectly with community members. Without this sup- knowledge demonstrates that you actually care port, the community members are often unwilling about the community, otherwise the commu- to participate or do not trust the new arrivals. An- nity is not likely to trust your motivations. Key other route into the ªeld may be working through words are “safe,” “communication,” and existing established relationships—either through “credibility.” family relations or through NGOs that are already • Critical Action Research may be a useful per- working on the ground (Medhi, 2007; Parikh & spective. Schwartzman, 2007). Communities are often suspi- • There has to be a sense of mutual learning— cious of outsiders coming in with big promises that the designers must expect to learn from the

10 Information Technologies and International Development DEARDEN

community and vice versa. This learning needs ference series, or a track within an existing to take place both before and during the proj- conference. A proposal for a track on interac- ect. tive for development has been • Sustainability needs to be built into thinking put forward for the HCI International 2009 from the beginning. conference. ■ • Evaluation and publishing papers has to be a Workshop Papers jointly owned activity. Just as with ethnog- raphers, we need to work out our ethical The full set of workshop papers can be downloaded stance in relation to the communities we work from: http://www.cc.gatech.edu/~mikeb/ with. We as researchers are beneªtting when UCDandIDWorkshop/accepted.html we publish papers; the community may not see Akorli, F. K. (2007, April). Sustaining a post-graduate any beneªt, but they do see a cost—they are study program in ICT in a developing country: A being represented. We need to ensure their case study at the National University of Rwanda. rights to fair reporting and involvement in how their story is told. Arora, S., & Joseph, P. (2007, April). Position paper: For participation in user-centered design and in- Looking Ahead ternational development. The ªnal part of the workshop was devoted to iden- Athavankar, U. (2007, April). Rediscovering UCD4 tifying key challenges and forward planning. In tak- development. ing our vision of community-centered design Aykin, N. (2007, April). Global innovative design for forward, some critical debates are required around social change. ideas such as development, participation, and com- munity. These ideas will strongly inºuence the gen- Bednarik, R., Kamppuri, M., Tedre, M., & eration of new approaches, new ideas, and new Vesisenaho, M. (2007, April). Alternative to tech- design contributions to development. Without such nology-driven development: An approach based a grounding, interventions may be well-meaning, on authentic needs. but ineffective or even counterproductive. Boettiger, S. (2007, April). Small Scale Agricultural Following are practical steps that were identiªed Technologies (SSAT) information resource. to support that debate: • Publishing some form of handbook, or text- Buckingham-Shum, S. (2007, April). Mapping dia- book, for ªeld practitioners and students. A logue and argumentation in international devel- ªrst step in this direction could be preceded by opment: The case of Compendium and introducing a chapter on interactive systems OpenLearn LabSpace. design in existing development studies texts. Chetty, M., & Grinter, R. (2007, April). HCI4D: How • Establishing and supporting networks of practi- do we design for the Global South? tioners and academics working in interaction Dhewa, C. (2007, April). Participation is like a jour- design in developing nations (Smith & ney. Dunckley, 2007; Aykin, 2007; Akorli, 2007). • Building international communities of collabo- Dickelman, G., & Arias, S. (2007, April). Fast and ef- ration that can discuss and explore relevant is- fective contextual analysis for ICTs in education: sues and developing bilateral links between User-centered design ªlls the gap. research groups. Diehl, J. C. (2007, April). INTOCONTEXT: Creating • Conducting follow-up on workshops and dis- appropriate solutions for peer to peer micro- cussion panels. Discussions were planned for ªnance in East Africa. HCI International 2007, Interact 2007, HCI Disse, K. (2007, April). UCD4D DIVO project—Dia- 2007, CHI 2008 and Designing Interactive Sys- logue for development. tems 2008. In the medium term, a structured effort is required, perhaps establishing a con-

Volume 4, Number 3, Spring/Summer 2008 11 USER-CENTERED DESIGN CONSIDERED HARMFUL

Frohlich, D. (2007, April). Applying digital storytell- Ouko, L. (2007, April). Designing for the DrumNet ing technology to community radio in India. application: A cell phone based information and transaction platform for Kenya’s agricultural sec- Godjo, T. (2007, April). Innovative techniques for en- tor. gaging users in food processing equipment de- sign in developing world contexts: The case of Parikh, T., & Brunskill, E. (2007, April). Lessons from Benin. prototyping a microªnance distance learning tool. Gupta, S. (2007, April). of SafalFin2 and user-centered design perspective. Parikh, T., & Schwartzman, Y. (2007, April). Estab- lishing relationships for designing rural informa- Hettiarachchi, C. (2007, April). Role of information tion systems. and communication technologies (ICTs) in human development in South Asia. Sharma, C. (2007, April). Background report on the Chikankari Computer-Aided Design (CAD) soft- Ho, M., Luk, R., & Aoki, P. M. (2007, April). Ap- ware. plying user-centered design to telemedicine in Africa. Smith, A., & Dunckley, L. (2007, April). Issues for hu- man-computer interaction in developing coun- Kam, M. & Ramachandran, D. (2007, April). MILLEE: tries. Mobile and immersive learning for literacy in emerging economies. Spittles, Y., & Dunckley, L. (2007, April). Ethical con- siderations for usability testing in Kenya using Kavanaugh, A., Reese, D. D., & Pérez-Quiones, M. Kiswahili and English. (2007, April). Mobile phone as scaffolding tech- nology: How low literacy groups learn comput- Toyama, K. (2007, April). Resolving user contradic- ing? tions through ªeldwork. Kitner, K., Beckwith, R. R., & Ilahiane, H. (2007, Wilson, I. B. (2007, April). Creating solutions that April). Development or stasis: The dilemma of are sustainable in context (case study of Sierra ICT4D. Leone). Kolko, B., & Rose, E. (2007, April). Design for Digital Winschiers, H. (2007, April). A cultural appropriation Inclusion research group. of usability engineering and participatory design practices—A Namibian experience. Kotze, P. (2007, April). User-centred design in a de- veloping country and international context. Wong, C. Y. (2007, April). Addressing challenges in bridging the digital divide and user-centered de- Liu, Z., & Meng, Q. (2007, April). Usability and digi- sign initiatives: A Malaysia perspective. tal divide: A study on usability issues for off-farm workers in China. Yeo, A. W., Songan, P., & Hamid, K. A. (2007, April). Providing equal access to ICTs for rural communi- Marsden, G. (2007, April). Position paper. ties: A user-centred design perspective. Maunder, A. J. (2007, April). Designing ICT based Yuan, Y. (2007, April). Get the right participants for solutions for the developing world. UCD studies in different markets. Medhi, I. (2007, April). CHI workshop position paper on user-centered design and international devel- opment.

12 Information Technologies and International Development