Notes on the New Design Space

Philip E. Agre
Department of Information Studies
University of California, Los Angeles
Los Angeles, California 90095-1520

This is a draft. Please do not quote from it.
Version of 2 September 2000.
4000 words.


Abstract. The design and use of information technology is increasingly embedded in the workings of institutions: the institutional order of the user community's collective life and the market order of the computer industry. As innovations in the basic technology greatly expand the space of potential designs, the design process must increasingly engage with this institutional context. This article is a meditation on the meanings of design and community in this new environment.


Information and communications technologies don't create much that's new; rather, people use them to amplify forces that already exist [1]. A hundred forces operate in society, and many of these forces conflict with one another. It follows that we cannot generalize about, for example, "what the Internet does". Nobody knows enough to evaluate all of the changes, and so nobody can tell us the bottom line.

Design, then, means selective amplification -- amplification, that is, we hope, of things we value. What are the technologies for amplifying community? Some communities operate at a low level because their members' lives don't bring them together enough, or because they don't have enough shared sense of identity. Yet within every community is a force toward a higher level of community life. A community needs a shared identity, a collective memory, a repertoire of ways of doing things together, familiar genres of communication, ways of moving along from newcomer to oldtimer, places and landmarks, rituals, a language and a songbook [2]. Information and communication technologies can amplify those things [3], or it provide some of their conditions, but in the end it's the community that does them. Design has to be okay with that.

The design space is exploding, and so design must change. Information technology has few opinions of its own. It is plastic, malleable [4]. Design means reconciling constraints, but now fewer of the constraints are dictated by intrinsic properties of the technology. Computing devices need no longer be heavy, or to sit in one place, or to be connected by wires. They can be woven into the artifacts and patterns of daily life in an unbounded variety of ways [5]. So we have to be more imaginative. The technology can take an infinite variety of forms, so we need more information before we can choose. We have to imagine our lives. Too much imagination is gadget-centered, or else driven by archetypes that are completely removed from the practicalities of the machinery.

Society is made of institutions: settled patterns of roles and rules, relationships and languages, repertoires and terrains of action into which we are all socialized [6]. We take for granted the workings of banks, schools, markets, meetings, the news, street traffic, visits to the doctor, and a thousand other arrangements. The weirdness and brilliance of information technology is that all of those arrangements are liquifying. Many of them existed to solve informational problems, or were limited by informational problems. They depended on a stable assignment of activities to places. But with information technology they are all suddenly being renegotiated. People used to drop acid to get this effect. But dropping acid is insanely dangerous, and now we can get the same effect for free. We can best see what a thing is when it's changing, and now everything is changing. The rules of every institution of society are being renegotiated, and we are either party to the negotiations or we aren't.

Design is increasingly public. One designs information technology by inscribing social discourses into machinery [7], and the machinery and the discourse coevolve. Technical standards need not embody a consensus in any formal sense of the word, but they certainly provide an outcome to a debate. Right now, for example, we are all oppressed by the poor to nonexistent model of human relationships that is inscribed in the software of personal computers. People who write viruses do a public service by making this fact visible. Will the public debate over such events cut deep enough to shape a new understanding of trust? Or will it simply lead to reactive lawmaking? We're growing a whole branch of government to protect us from the poor model of society that's embedded in Microsoft products. That's not good.

What is public design? Hackers, in the original sense of the word, change the world by designing things. They see the power of a tool. They value rough consensus and running code. But intellectuals are hackers too. They intervene in the collective cognition of a society. These are grave responsibilities. Ideas become machinery; machinery becomes ideas. Each can set the other straight. Each can confuse.

E. P. Thompson wrote an influential paper about the fashion for pocket watches when pocket watches were new [8]. People would spend large parts of their wealth to get one. The watches coordinated activity, and they also signified. Life by the clock was new then, and it was also news, fashion, as the culture tried to figure it out. Neither happens without the other. What does the Internet signify? The Internet came along at the moment when engineers -- at least the leading cultural edge of them -- were surrendering the centuries-old conception of engineering as the Godlike giving of rational order. What's to take the place of that conception? Engineers facilitate local ordering. Rather than discover an optimal order, they provide a platform [9]. That's the ideal, which we code with the word "open". But platforms, alas, are public goods. If markets epitomize the otherwise inchoate wish for an open society, we've been learning that marketplaces are commodities too. And the market in marketplaces is, very often, a natural monopoly.

As society liquifies, design becomes a way of life. The word design has been understood in two broad ways, deriving from engineering and from architecture. The two understandings are converging. It was once thought that engineering is a rational process. But we've come to a deeper understanding of process. The architects were right: design is a process of discovery [10]. Can engineers get used to discovering things? If you accept the possibility and necessity of discovery then you surrender the expectation of knowing everything. You accept that you are not God. Engineering reason still has a purpose, though. One creates temporary islands of order, discovering which abstractions might be broadly useful to others and then rationally designing those.

As platforms multiply, it becomes possible to do amazing things in an afternoon, in a garage. The platform provides more and more of the functionality. That, again, is the ideal. Some people use the term "platform" to refer to programmable hardware, but the deeper meaning is: something that you can build things on top of. Platforms nest, with new ones built on old ones. As the platforms stack up, the newer ones come closer to the experience and concepts of ordinary people. A microprocessor does little that any ordinary person cares about, but a payment system, for example, needs to speak the same language as the people who use it. Platforms become geologic strata, rigidly interwoven with a diversity of practical arrangements in society, so it's important to get them right.

The rising tide of standardized platforms presents a challenge for design. Good design means having a single, clear concept that can inform the design on every level. The internal architecture of the machinery, the packaging it comes in, its physical form, the ways you get help with it, the community of people who use it -- all should express the same idea. But the design space is exploding because of the economies of scope that come with standardized platforms. A new conception of the world can only cut so deep into a design if seven-eighths of the design are ordered from a catalog. Traditional "systems analysis and design" classes imagine a world of bespoke software, every line designed from scratch. The world isn't like that anymore, if it ever was. The design of information technology is collective and cumulative. One isn't designing for a green field but a tangled network.

The concept of cyberspace is an artefact of old-fashioned technology. The "computer terminal", with its big flat screen, presents a membrane between two worlds, often glossed crudely as "atoms and bits" [11]. It's not like that. We're going to redistribute information technology so that it's woven into our lives. The idea that information technology is a place apart from the rest of the world was always just a lobbying campaign: your laws don't apply here. What a pallid form of imagination. We will not migrate into a realm of bits. Rather, we will reorganize and redistribute our lives, both individually and collectively. We will codesign our technologies and our ways of life. We can do this well, or we can do it badly.

Who is "we"? The collective design of information technology is a complex thing. Here's a simple model. Will we move to technology A or technology B?, we'll all ask. Both technologies exhibit network effects, so that they are only useful if a critical mass of other people is using them [12]. And everybody knows this, too, having seen it happen with other technologies previously. So even though the two technologies might have real benefits and disadvantages for various groups, nobody will move until they figure that everybody else is going to move. The reflexivity of this process is exquisite. The supposedly decentralized society of information technology is always building consensus about design issues that affect everyone. It is nothing short of a political process. Self-fulfilling prophesies will count for a lot. So will the forms of imagination that, when shared, make consensus easier if not automatic.

It's not really a consensus, however. That suggests that everyone's voice counts equally. Network effects need a critical mass, not a unanimous agreement. The parliament of technology enfranchises the early adopters, the ones who create (or fail to create) the critical mass for a new technology. And the early adopters are the affluent and the gadget-obsessed who value the devices for themselves and not what they can do. We need forms of imagination that are anti-gadget.

Professional designers, perhaps, are latecomers to the scene of the accident. The real design has been done by poets, and intellectuals, and propagandists. And by social movements. Designers work with the raw materials of form and meaning. Sometimes they are arbitrageurs. To the extent that the design of information technology is collective, design becomes a way of life. You can't just move from job to job. Standards are designed by committee, and they build network effects in the technical public sphere. Design becomes advocacy. The designer becomes a representative for all of the people whose attentions are still elsewhere, who don't know the stakes in a design process whose results will become irreversible by the time they ever hear about it.

What would it be like to have amplified communities? Think about the people whose role-playing games can involve dozens of people and go on for weeks. They can remap a whole landscape with alternate meanings and become lost in the drama of their alternate world. That is one model. It's immersive in the truest sense: not being enclosed in a box of machinery, but imagining one's way into an alternate terrain [13]. One can be enclosed in a box, surrounded by high-resolution displays, and not be immersed at all. What would an immersive artwork be like that one would want to inhabit for 24 hours?

Media companies merge in order to provide immersion in the broader sense. A child can live all Pokemon all the time -- Pokemon in every medium, through every sense, every moment. From an economic point of view, the media company is achieving economies of scope. The work that goes into devising the symbols -- the characters, narratives, design themes -- is leveraged across more products, and the work done on each can promote the others. Someone who is immersed in Pokemon can find numerous outlets for that immersion. Pokemon stands out because of the uncanny sense that small children live in a different world that is increasingly colonized by large media companies, but the same goes for a lot of other synthetic meaning systems as well.

Social movements can be immersive. Indeed they probably do not succeed unless they are. A social movement needs cultural forms, and it needs institutions. It needs to provide everyone with something to do, and it needs to spot and recruit talented people. It needs to provide a repertoire of action forms, and this includes both ways of changing laws and ways of having a party. The challenge, of course, is to be all-embracing without being oppressive. Social movements thus rise and fall, each becoming rigid and providing fuel for its opponents. This is healthy, perhaps, in the very longest run, wasteful and often dangerous as it seems in the short run.

What makes a community worth an investment of one's time and effort? The strength, in some spiritual sense, of the people who are involved in it. The degree of shared meaning that they can build up without turning that shared meaning into a way of controlling one another. The distribution of labor: democratic organizations are too often run by a few people who carry the entire load and then burn out or become resentful. The quality of leadership: framing a vision rather than manipulating. The culture: people either have good habits or they do not, and they approach their joint activities from either a standpoint of positive expectations or a standpoint of powerlessness. Could a machine ever fix such problems? It's very much the wrong question. If a community and its machines are both informed by a common vision, then other communities that don't work can fade away.

Here's the paradox of design: design is supposed to make something new, but it depends on astute observation of what already exists. Many technology designers resist observation because it feels conservative: who cares what already exists, given that we're about to make it obsolete? And indeed, observation can be shallow. When you look hard enough, you see just how interconnected the elements of our existing world really are. How could one possibly change it? Yet it does change. "People resist change", the control freaks say, but this is a libel. What they mean is "people resist my changes". In truth, everyone embraces change -- longs for it. So what is the role of observation? Observation must be analytical, seeing the world not as a tangled whole but as a process in motion, a surface with depths, with moving parts, with forces whose development has been stunted so far by the means readily available to hand. What forces do we want to equip? That's the question for design. People will embrace changes they can understand, and so one must observe understandings and engage with them. If designers are no longer the engineers of human souls, what can they be? They can be interlocutors, partners, provocateurs, even leaders, just so long as they comprehend their own location in the whole.

The new design space afforded by information technology is basically about connection. Computing power is only somewhat useful by itself. It's good for speech recognition and a few other things. But the technology mostly challenges us to become aware of our embedding in a set of relationships. This can take a lot of forms, which we can explore and reinvent. Here's a thought experiment. Every one of us has six billion relationships -- one for every other person on earth. Each of those relationships has its architecture: you are near or close, you disclose some information and not other, you possess some information and not other, you are on good terms or bad, speaking or not, you are doing business, and your relationship is more generally embedded in one or more of a thousand institutions with their rules and expectations. You must manage every one of those six billion relationships, and in a networked world you must increasingly manage every one of them in real time. Once you would call a friend once a week; now many people call one another ten times a day. Why not a hundred times? The calls are shorter, more incremental.

The thought experiment is extreme, of course: we are finite beings, and we could never keep track of six billion others in real time. But professionals often keep track of thousands of people in their social networks, and more generally in the universe of people whose careers they track as part of their own institutionally organized world. Agrarian societies, perhaps, had no use for a concept such as "networking": having and consciously applying a conceptual framework for initiating, evolving, and managing large numbers of relationships in a strategic way. The very word "networking" retains the nasty connotation of "knowing people" and "politics" -- a sense that rich people and managers just build social networks rather than doing real work, and a sense that someone who moves dextrously among many people can mislead them all. In a dynamic world, however, networking is a fact of life. Networking skills need not be organized by means of a conscious framework of concepts and strategies, however. The people who live immersed in new communications technologies, and adolescents in particular who invent new customs around the technologies as they set about inventing themselves, are pioneers of new cultural forms for managing relationships. The technologies themselves presuppose overly simple or outright false models of these relationships. What other models could they presuppose?

The standard computer-science approach to answering this question is similar to that of the standard business-management approach: make a conceptual framework, then embed that conceptual framework in practice. But as computer systems are increasingly intertwined with real life, this design practice is showing strains. A whole research community is concerned with "groupware" [14], and the central problem of groupware is that the systems are too rigid. One frames a concept such as a "task" or a "role" or a "rationale", builds data structures that go by those names, and expects people to represent themselves to the machine in those terms. One then discovers that group activities are much more complicated than that. For computer science it's a puzzle: groups speak of tasks, roles, rationales, and so on, and yet they are not able to represent those things digitally.

This is the problem of structure in system design, and it is a hidden crisis. One response is to discipline everyone to conform to the structures. Once the flow of work has been inscribed in the machine, everyone has to organize their work in the way that the machine says. Life becomes more structured, and this increase in structure may be the point. It's a means of control in senses that can be both bad and good, depending on where you stand. Another approach is to relax the demands for structure. Paper documents, for example, are typically semi-structured. Even the most rigid form provides a lot of room for negotiation, and document genres exist largely to be violated, and not as grammars to be followed in a mechanical way. Document genres create expectations, but they do not necessarily impose constraints [15]. Computer science assumes that all structures should be formalized, since after all a computer can only compute with data that it can capture, and captured data is only meaningful within a structure of standardized meanings and practices [16]. Data can only be searched, aggregated, and computed with to the extent that it is structured. Because it is concerned almost entirely with the inside of the machine, computer science has not reckoned adequately with the deep trade-offs around structured data. And because of this, as Judith Gregory observes in her new dissertation [17], computer science is utopian in a profound and not especially useful way.

No wonder, then, that the rest of the world has come to see computers as a communications medium. The price of structuring one's life in order to create structured data is too high, but the benefits of being able to connect with others electronically are enormous. We are still using very unstructured email and voice communications, and we do not submit our electronic interactions to all that much structuring. The best "community" systems do both: imposing small amounts of structure on the interaction but then providing free-text interaction forums as well. EBay [18], for example, imposes moderate amounts of structure on the auctions that users post, but leaves a big space for unstructured text describing the goods. It offers structured choices about payment terms, but also allows users to opt out of those structured choices and simply explain the terms in the unstructured text. It imposes structure on the bidding process, emphasizes that the structure derives from legal contracts, but then provides limited room for negotiating one's way out of those contracts. It counts "feedback" points but leaves room for free-text feedback comments and replies up to eighty characters. It supplements the structured auction process with relatively structured forums. At each point, the border between structure and free text is thought through, or evolved. To computer science, however, this picture will be frustrating, because for computer science progress depends on discovering and imposing more and more structure on data. Can't the world progress beyond eBay? To the computer science way of thinking, it's a sad thought.

Still, eBay is a "community" in only the most primitive sense. Being a marketplace, it really does push toward the extreme "six billion relationships" model of society. Everyone's goods are on display for the world to see, and everyone has a public persona. Economic sociology makes itclear that markets are embedded in societies much more broadly, but this embedding can be structured in many ways [19]. Embedded or not, some market mechanisms are more impersonal than others. Our conceptions of "market" and "community" are in tension no matter what the economic sociologists say, and one can have very different sorts of "communities" on different scales. The world's stock exchanges are embedded in a community as well, one drawn together by the financial press and the democratic public sphere more generally. But this community is no longer a small world. And what is eBay is embedded in? The answer will presumably be different in each distinct market that eBay mediates: the antiques world will work differently from the computer world or from consumer electronics. Standardized products can be traded more anonymously, for example, and markets in nonstandardized products demand greater knowledge. In some markets vendors often buy from one another, and in others a customer will develop long-term relationships with a vendor who then buys for that customer on the open market.

What does design mean in such a world? Design needs the conceptual equipment to draw deep distinctions among social worlds, and to notice deep analogies between them. When activities and relationships are structured so deeply by institutions, and by the particularities of communities and their customs, and by the diverse practicalities of embodied activities, one cannot design for "people" in general, much less for "users". Nor can one expect to design a gadget that's useful on its own. Just as participation in a social world is a long-term immersion, the work of the designer is long-term as well. Technology is indeed moving quickly, but the capacity of social arrangements to digest technology will always be limited. The huge quantitative improvement of the technology provides an irresistible force, but the social world provides an immovable object to meet it. Not immovable, exactly, but movable on its own terms, along its own vectors, with deep respect to its own forms of imagination and its own structure of interests. Design can participate in these movements. It can lead them, in the deep sense in which leadership, as a variety of politics, is the art of the possible. Or it can follow them, or confuse them, or become irrelevant to them. As life itself becomes inescapably a matter of design, design necessarily becomes continuous with life, or coterminous, as the case may be.


I appreciate helpful comments by John Chris Jones and Cosma Shalizi.


1 Dutton, W H Society on the Line: Information Politics in the Digital Age Oxford University Press, Oxford (1998)

2 Wenger, E Communities of Practice: Learning, Meaning and Identity Cambridge University Press, Cambridge (1998)

3 Agre, P 'Designing genres for new media' in S Jones (ed) CyberSociety 2.0: Revisiting CMC and Community Sage, Newbury Park, CA (1998)

4 Danziger, J N, Dutton, W H, Kling, R, and Kraemer, K L Computers and Politics: High Technology in American Local Governments Columbia University Press, New York (1982)

5 Norman, D A The Invisible Computer: Why Good Products Can Fail, the Personal Computer Is So Complex, and Information Appliances Are the Solution MIT Press, Cambridge (1998)

6 DiMaggio, P J and Powell, W W 'The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields' in W W Powell and P J DiMaggio (eds) The New Institutionalism in Organizational Analysis University of Chicago Press, Chicago (1991)

7 Friedman, B (ed) Human Values and the Design of Computer Technology Cambridge University Press, Cambridge (1997)

8 Thompson, E P 'Time, work-discipline and industrial capitalism' Past and Present Vol 38 (1967) pp 56-97.

9 Messerschmitt, D G Networked Applications: A Guide to the New Computing Infrastructure Morgan Kaufmann, San Francisco (1999)

10 Schon, D A The Reflective Practitioner: How Professionals Think in Action Basic Books, New York (1983)

11 Mitchell, W J City of Bits: Space, Place, and the Infobahn MIT Press, Cambridge (1995)

12 Shapiro, C and Varian, H Information Rules: A Strategic Guide to the Network Economy Harvard Business School Press, Boston (1998)

13 Moser, M A and MacLeod, D (eds) Immersed in Technology: Art and Virtual Environments MIT Press, Cambridge (1996)

14 Baecker, R M (ed) Readings in Groupware and Computer-Supported Cooperative Work: Assisting Human-Human Collaboration Morgan Kaufmann, San Mateo, CA (1993)

15 Yates, J and Orlikowski, W J 'Genres of organizational communication: A structurational approach to studying communication and media' Academy of Management Review Vol 17 (1992) pp 299-326

16 Agre, P E 'Surveillance and capture: Two models of privacy' The Information Society Vol 10 (1994) pp 101-127

17 Gregory, J Sorcerer's Apprentice: Creating the Electronic Health Record, Re-inventing Medical Records and Patient Care Doctoral dissertation, UCSD Department of Communication (2000)


19 Polanyi, K 'The economy as an instituted process' in K Polanyi, C M Arensberg, and H W Pearson (eds) Trade and Market in the Early Empires: Economies in History and Theory Free Press, Glencoe, IL (1957)