Imagine having a device which can post sound bites from your home to your Facebook wall. Now and then it will record the sounds in your home, scramble them – so listeners cannot recognize specific sounds, and it will make an audio post for you. Most people I talk to, think this is a stupid idea. But it is what “Facebook Listener”, designed by Stefan Veen and his colleagues, does – and I do believe it could be a success. Or at least I believe in the broader idea of awareness systems: allowing people to share background information with others in a lightweight way. In this post I will answer four critical questions I often get about awareness systems and I hope to show where the opportunities for awareness are.
A short history of awareness
Maybe I should to take a step back and discuss shortly where some of the ideas behind Facebook Listener originated. The academic tradition of awareness systems started in the workplace of the early 1990ies. When people collaborate they build on tacit knowledge about each other. Several workplace studies showed that people who have an awareness of their co-workers can collaborate more easily and effectively. Scientist turned out to publish most articles with people whose office was nearby. People in control rooms needed details about what others were doing to avoid mistakes. Because of findings like this, researchers slowly started to recognize a design opportunity: “If people need background information about each other to work together”, they thought, “can’t we design systems which deliver such information across a distance?”. So soon, researchers started to build devices which allowed awareness, background contact and casual communication at a distance. Mostly these media spaces, as they were called, involved video links between multiple remote offices. This was exciting and controversial work: in those times computers had to be useful tools to complete specific tasks effectively. The idea to use computers to provide background information about others seemed to come from outer space. Exciting or not, researchers needed several incarnations to get the design of these systems right, and honestly, the early controversy didn’t fade.
Is there a need for awareness systems?
The top 1 question. Does your work, for example, improve if you get a video connection with remote colleagues; so you can see them behind their desk, typing? Or, does anyone benefit from listening to scrambled sounds from your home on Facebook? Perhaps not: I have not seen studies of media spaces which led to measurable improvements on business results and clearly my friends survive without being able to listen to my house sounds today. But, probably yes too. Consider Digital Family Portrait, for example. Researchers from Georgia Institute for Technology placed sensors in the home of an elderly woman so the computer system could get a sense of her activity during the day. The (adult) children could see this activity data on a digital photo frame in their home. There were butterflies around the picture of the elderly woman and big butterflies meant “mother is active”, while little butterflies pointed to the opposite. Digital Family Portrait raises many questions, but the users were positive about it and it shows how awareness systems can support real communication needs.
Often, our communication is not so much about the contents of the communication but for the sake of communicating itself. This is called phatic communication. When I share what I am eating on Twitter, my followers learn something boring and something important about me. The boring part is what I am eating. The important part is that I am alive, all is well, and the communication channel is open. They can contact me if they want to. It turns out that this last phatic part of the message is important, in particular between close ones. The users of digital family portrait didn’t really need to know how active their grandmother was. But they needed a possibility to check on her regularly for their peace of mind. Digital Family Portrait provided them an easy way to do just that. There are other reasons the butterflies on Digital Family Portrait made good sense as well. The meaning of a bit of information about someone close to you depends on your background. To the users of the Family Portrait the size of the butterfly is not so important, but the way it changes over time is. A sudden drop in activity, for example, means there is a reason to check on grandma (or the hardware). Finally, background information such as Facebook Listener, Media Spaces or Digital Family Portrait give, can be a starting point for a more intimate and meaningful conversation. Users of Digital Family Portrait could use the butterflies to talk about health and lifestyle choices. So it could be that Facebook Listener, Digital Family Portrait and Media Spaces aren’t presenting the right information about the right people in the right way, to the right people. But this doesn’t mean the general idea is wrong. It seems people want to share and receive background information about others. Even if it is just phatic communication, I would say there is a need for awareness systems.
Why spyness systems? What is wrong with human updates? Aren’t those sufficient?
Granted that people need background information from others and that they already share such information on social media, why bother any further? Or stronger: aren’t we crossing a hard and scary border when we start to capture and send information about people automatically, like the three examples so far do? I find this question much more difficult to answer, but I would say no: it fine to build systems which give automatic updates. Perhaps JK Rowling can support my argument. In the Harry Potter book “The Goblet of Fire”, the Weasley Family has a clock which shows were family members are, rather than the time. This is an example of automatic updates that seem sympathetic and useful. In fact we know from research it is. Soon, this “whereabouts clock” as researchers called it, was build and evaluated by Microsoft Research . They report that – at least within nuclear families – the clock gave a sense of reassurance, connectedness, expression of identity and social touch. The clock quickly became an integral part of the routines of the family members: it provided phatic communication possibilities. Family members got a sense that everything was going to routine, that all was well. Part of the success of the clock may have been its Potterish design, though. Much like the original version of the Weasley family, and unlike some commercial tracking systems, Microsoft’s whereabouts clock used rough place labels like ‘home’, ‘work’, ‘school’ and ‘elsewhere’, rather than precise GPS data. Probably there is a tradeoff in automatic sharing. The more intimate the data you share gets, the more abstract, course grained or fussy the presentation of this date needs to be, to be acceptable to users.
Who needs even more ‘information display’s’?
This question comes in different forms, but the gist is that people wonder how to act on the new information they are getting. Wouldn’t it be awkward to pick up the phone and say “why’s your butterfly so small today ma”? The notion of awareness is information centric much more than communication centric indeed. Many awareness systems researchers want to figure out what people could share though sensors and how to show this information to users. They seem to care less about the next step: providing the means for the ‘talk’ that may follow from the awareness. Often, like in Digital Family Portrait the information stream is one directional too. The mother lacks common ground, she doesn’t know how big her butterfly is that day, which creates the awkwardness of the example. Some researchers argue that people have enough channels and the awareness system only needs to provide the ‘trigger’ for the communication. But I think the evidence from the studies I mentioned points in a different direction. Users of Facebook Listener turned out to use it to create dedicated ambient-home-sound-scape messages for their friends. Microsoft researchers reported that users of the whereabouts clock wanted to be able to send messages back to their peers from the clock. Evaluators of other awareness systems report similar findings. Awareness information triggers communication needs, and it seems normal to support those from the system, preferably in an open ended and playful way.
Who would buy such a system?
This is a fair question. A positive evaluation doesn’t form a business case. There are two serious barriers to market entry of current awareness systems: single purpose systems and the network effect. Many awareness systems serve a single purpose: a specific awareness need or a single type of awareness information. This fine for research projects, but it creates small markets and high entry costs. It seems difficult to overcome. Users of the Digital Family Portrait and the whereabouts clock can understand the information of the system because the design supports the use case. Some have tried to create more flexible awareness systems such as the modestly successful awareness rabbit called Nabaztag, but this is difficult. It is a pity, but Nabaztag cannot display location information of the family in an optimal way. When we abstract away from (design for) a specific type of information the awarness display becomes hard to read. A related problem is the network effect: part of the value of any communication system is in the number of users that already use it. Awareness systems, usually being specific purpose, a-symmetrical and standalone, may suffer from this effect. The decision to buy or use is not with a single person but at least with a small group. This makes those awareness systems that make use of existing technical infrastructure and use patterns (of mobile phones for example) much more likely to succeed than other sensible awareness proposals.
So the proposal of Stefan Veen and his colleagues: to make awareness systems which cooperate with Facebook makes a lot of sense from a market perspective. The idea is resembles our proposal of a year earlier: to integrate social media and business by building software on top of the existing social media (integration software). The benefit could be that users are already networked in this environment and they can make use of its lightweight communication tools. But this benefit comes with two costs. First, there is a usability challenge. In our own research we found that it is hard to communicate to users how the use of the new interface works together with Facebook. In a way users handle two systems at the same time. It is not easy to understand how an action in the integration software or awareness system plug-in changes the state of Facebook, for example their timeline. Second, Facebook integration only has an advantage if whatever comes out of the awareness system is also informative for non-users of the system. Any message the new system puts on Facebook needs to be a good addition to the facebook timeline. This is a strong limitation. So while I expect that awareness systems will be build that will integrate with social networking sites, chances are that they will most likely provide a new output of the data which is on Facebook already, or can be inferred from it, rather than providing a new input device. But on the other hand, if awareness systems are eventually communication systems new inputs will soon be added to this outlet in your home.
Much of the information for this blog post comes from the book “Awareness Systems: Advances in Theory, Methodology and Design”. In particular: the history of awareness (chapter 1), phatic communication (chapter 7) and the wheraboutsclock (chapter 18). The design of Facebook Listener is described in detail in this NordiChi paper, the Digital Family Portrait in this paper. Our exploration of software which uses social media as an infrastructure for new applications can be found here.
 Several incarnations where build at Eindhoven University of Technology as well.
Filed under: (re)thinking media, discussion, review | 3 Comments
Tags: Ambient Intimacy, Awareness Systems, Common Ground, Digital Family Portrait, Facebook Listener, Media Spaces, Nabaztag, Phatic Communication, User Experience
A book review
It is not an easy challenge Lev Manovich sets in his book “The Language of New Media“. In the early days of cinema hardly anyone could foresee the enormous cultural impact this new medium would have on our society. A new artistic language, cinematography, was born; but no one recorded its first steps systematically. This wasn’t only because people didn’t see the importance of cinema. Documenting an emerging language is a form of historiography; merely impossible without the advantage of hindsight. It simply takes time, before it becomes clear what the right or interesting historical questions are. So, I never believed Manovich, could succeed in “providing a potential map of what the field could be” (p11) back in 2001. But, I do feel his book deserves a close reading. The first steps in building a theory are usually the hardest and, despite the difficulty of the task, Manovich did cover interesting ground.
Manovich tries to understand new media with the eyes of an artist. Two ingredients are necessary for every art piece. First there is the influence of existing media. A key concept in Manovich’s theory of new media is the cultural interface (p 69). Much in the way the human computer interface, structures the interaction between humans and the computer, cultural interfaces give structure to the users’ interaction with culture (or “cultural data”). The interfaces of CD-ROMs, web pages, games and apps are all cultural interfaces. Often new media reuse ideas, forms and conventions from older media. Early cinema build on theater, rock music on blues. So the forms of new cultural interfaces also stem from older, already familiar forms such as: magazines, newspapers, photography or -indeed- cinema. Second, the new media have offer new technological possibilities and affordances to the artist (Manovich speaks of operations). As media makers experiment with these new operations, some older conventions will fade and new conventions emerge. So to understand the language of new media we need to look carefully at its ancestors and at the possibilities of computers.
Manovich believes three older cultural forms are most important to describe the language of new media: print, cinema and the human computer interface (in practice Manovich specifically refers to the graphical user interface, the GUI). Print, the oldest form, was adopted first. The page is a cultural convention of print that persisted into the digital age; although the worldwide web also revived the ancient form of scrolls. Hyperlinks were the most radical and disruptive innovation to texts. Hyperlinks challenge old ways of organizing information. The structure of the web isn’t like a library (with an index) or a book (structured through rhetoric narrative) (p 77), but more like a walk on the beach where you may find random objects one after the other (p 78). Cinema, is the second cultural form that influences new media; in particular games. The moving camera is a convention borrowed from cinema, apt for navigating in virtual worlds. Also many games borrow story forms from cinema. But games moved beyond cinematic conventions too. They break the rules of (natural) perspective and story, in search of forms that are more suitable for interaction. The graphical user interface (GUI) brings controls, menu structures, and the desktop metaphor to new media. These seem to fit in, but there is friction as well. The controls offered by GUI’s need to stick to the underlying metaphors (for usability) and simultaneously blend in with the story world of the cultural interface. While new media often show mixtures of print, cinema and graphical user interfaces these mixtures may be rough. Often, the underlying ideas of what the screen represents differs so much (flat surface with information, window into an immersive environment, control center), that these forms cannot be wedded easily.
So print, cinema and GUI’s, are to new media what photography and theater were for cinema. They provide the “raw material” of cultural conventions that are available to new media makers. But they do not yet describe the other major influence on the language of new media: what new media creators tend to do with this material. In other words, Manovich needs to turn to the affordances of new media creation technology, or as he calls it: the operations. Cinematography, for example, makes creative use of different types of shots and editing techniques. Do these operations have new media equivalents? Manovich believes these are selection, teleaction, and compositing. Teleaction, the ability to see and act at a distance, allows the camera to be everywhere and users to cooperate across the world. In massive online multiplayer games for instance or in massive online open courses. Computers allow easy access and reuse of older material and both selection and compositing capitalize on this possibility. The rise of the DJ, cleverly choosing and combining existing materials to create new music, is an example of the power of selection and compositing as means to create new forms. The growing importance of special effects in movies, another.
New media creation technology also allows to create new types of illusions. Computer generated images for example. Mimesis: to mimic nature, has been an important goal of cinema and this remains so in computer simulation. For computer vision scientist, for example, cinema is an important market and source of inspiration: “High quality, means virtually indistinguishable from live action motion picture photography” (p191). In practice, creating a realistic immersive experience involves more than just photo-realism. It involves many forms of mimesis like: touch, interaction (with virtual characters), moving about in virtual space and 3D graphics. But mimesis isn’t at all one-dimensional. Rather, in cinema, the quest for realism progressed through a succession of ‘codes’ in which only some parts of the experience mimic real life and the viewer filled in the gaps. In computer graphics ‘realism’ was first achieved by ‘deep perspective’ and later by ‘correct lightening and shading’. Considering the broad playing field for new mimetic codes in new media (for touch, interaction, movement and so on), we may expect a long period in which media makers try to set new mimetic frontiers.
After describing the conventions and affordances of the new technology, the cultural interfaces, operations and illusions, Manovichs turns to the emerging genres or forms. He focusses on two of those new forms: the database and navigable space. Databases are special because they have no beginning or end. Much more than any old form, new media objects allow the user random access to items in the piece. The user experiences new media through hyperlinks, browsing and searching rather than through the guided tours that traditional narrative forms offer. The second form is navigable space, which is the dominant form in many games and interactive stories. It is the first time space is a medium. Space can now be stored, retrieved, transferred and it can be used to tell a story. Today, architecture touches media and new media designers need to learn how to tell a story with space.
The Language of New Media, is a rich book which offers a comprehensive theory of new media. It is an interesting idea to use cinema as a model for new media and with this approach Manovich could get to some marked insights in the language of new media. But I do not feel the book can live up to its goal to be a map of the field. Manovich tries to bring clarity to the field by thinking of new media in layers: cultural interfaces, operations, illusions and forms. But the distinction between these layers is not always clear and he fails to show how these layers influence each other. How did, for example, spatial navigation as a form emerge through the operations selection and compositing on cultural interfaces? Manovich raised the question, but he gives no answer. There were many other questions like this, that haunted me when I was reading the book. At the end of the day Manovich book is more a structured description of the status quo of new media at his time, rather than a theory through which we can understand it’s development. And then there is the scope of the book…
When you compare (early) cinema and new media, you start inspecting the parts that are most like cinema. This way, Manovich misses a lot. As an interaction designer I can hardly agree with Manovich treatment of the human computer interface (and its history). Interaction design in games draws much more from board games and free play, than from desktop tools. So board games are important cultural interfaces that feed new media. Also, in contrast to film, new media has multiple context of use, and is paid with a bigger diversity of business models. This must have an impact on the developing language of new media, and studying the history of cinema cannot tell you what it is.
Shortly after the language of new media came out, web 2.0 unfolded. This changed what most people see as the dominant forms of new media. As early as 2005, four years after the language of new media was first published, Mark Deutze, considered “participation”, “remediation”, and “bricolage”, to be the principal ingredients of digital culture. Deutzes description of – in a word – participatory remix culture, would be part of any new book on the language of new media, but Manovich could not have foreseen this development when he wrote his book. Manovich knew that this might happen, as it happened with cinema before. He writes: “It is tempting to extend this parallel a little further and speculate wether this new language is already drawing closer to acquiring it’s final and stable form, just as film language acquired its classical form during the 1910′s. Or It may be that the 1990s are more like the 1890ies, in the sense that the computer-media language of the future will be entirely different from the one used today”.
So did Manovich write his book 10 years too early? Probably not, rather I think his scope was too wide for his analytical means. Cinema has never been the ‘meta-medium’ which the computer is today. Cinema, has always’s been expensive to produce and it has always offered a special and exclusive experience to the audience. So the technology and the economics (business models) of cinema quickly converged, in turn allowing for stable cinematic conventions to emerge. The meta-medium of the computer is cheaper, more diverse in the types of experiences it can offer, the contexts of consumption and the diversity of business models available to support their production. There is not one language of new media, there are many. And unlike with cinema, they are not likely to revolve around a common core any time soon. Manovich was right and observant when it came to games and digital cinema, but he may have underestimated the power of (linear) narrative, overestimated virtual reality and he may have missed forms like augmented reality or social media. Many new media forms, like immersive games and websites have stable language for years. For these forms, Manovich book provides at the least a descriptive framework. But other forms, such as augmented reality, social media and embedded media have their proverbial 1890ies still to come.
In my post Cognitive Bias in the Global Information Subway , I discuss the language of search and its impact. In Collateral Damage of the Robots Race (on the Web) and Social News Needs a Nuanced ‘Like’, Quickly, I discuss the impact of artificial intelligence on the way web experiences are structured.
Here is a good Dutch summary of “The Language of New Media”.
Filed under: (re)thinking media, review | 1 Comment
Tags: Book Review, Business Model, Cultural Interfaces, Lev Manovich, Semiotics, The Language of New Media, User Experience
I guess most design students have learned about integration at some point. Business, Costumer-relations, Engineering, Marketing and other departments may have very different demands for a new product and the designer has to find her (his) way across the different value sets and constraints of these different departments. An integrated design manages to do so; elegantly. Integration is difficult enough when the product is ‘just’ a new loot on a product family tree; but in innovative projects integration can be daunting. In this post I would like to address such projects and show how modeling ‘integration’ in innovative projects can relief some of the design tension that emerges in integrative innovative design projects.
In his PhD thesis Describing Design, A Comparison of Paradigms, Kees Dorst defines integration as follows: “Someone is designing in an integrated manner when he/she displays a reasoning process building up a network of decisions concerning a topic (part of the problem or solution), while taking account of different contexts (distinct ways of looking at the problem) or solution”. This is just a formal way of saying a designer needs to take different (incompatible) viewpoints into account in the design process. So: for innovative design projects the question is if there are generic viewpoints that apply to most projects and capture most of the design space. Finding such viewpoints takes experience, but for the interactive product design projects I worked in: user, design, and technology always were the ‘big three’. ‘Business’ may be a fourth, but I’ll get back to business near the end of this post.
Of course, I am not the first to highlight user, design and technology as the most important perspectives in interactive product design. In their seminal paper called “What Do Protoypes Protoype?”, Charles Houde & Stephany Hill make a distinction between four types of prototypes: role prototypes, implementation prototypes, look and feel prototypes and integration prototypes (see the figure below).
Each of these prototypes forms a tangible and temporary answer to a design question. Role prototypes answer the question “what changes in the life of the user because of the new product”. Look and feel prototypes are about the sensory experience of the product. Implementation prototypes address the question how the product will work. Finally, integration prototypes answer multiple of these design questions at the same time.
Rather than just a prototyping concern, the three corners of the Houde & Hill model are general concerns in innovative design projects. So to use the Houde & Hill model as a model for integration in innovation, you need to interpret it more broadly. Therefore we extended the model to include three relevant contexts for design and three innovation forces (see the figure below). The first innovation force is user pull. A design team shows to master user pull if it has a concern with the user and the utility of the product in the context of use. The team creates new user scenario’s in writing and with storyboards (role prototypes). Most of you will recognize this as a user centred design capability, but ‘pure’ UCD is not enough. If the other two forces: design push and technology push are neglected the result may not be innovative and may not appeal to users (beyond the utility of the product). A walker for example, fulfills all functional user needs but somehow something seems to be missing. Perhaps walkers do not appeal to soft values such as status and power because of a lack of design push in the design, and possibly the latest technologies could bring improvements to walker design too.
The second force is design push. For design push a team needs to be sensitive to the social and cultural context and have the ability to translate this into design solutions. Anthony Dunne’s critical design (design for debate) movement, aiming to expose undercurrents in society, is an example of design push in its purest form. But there are lighter forms of design push too. For example: many designers are able to translate brand values into a design, thus showing sensitivity of the social cultural context. In the Houde & Hill model, design push ends up in the ‘look and feel’ corner, but I do not like this term at all. A good ‘look and feel’ prototype expresses meaning in a specific social cultural context, which is more than a ‘pretty picture’. Therefore I prefer the term semantic prototype. Design push is a positive force, but there can be too much of it too: projects with too much design push are ‘interesting’ or ‘provocative’, but not always useful.
The last force is technology push which is the ability to identify new technological developments and to appropriate them for the design. Of the three forces technology push has the worst name. You do not need to be a user centred design fundamentalist to know of a product that is mostly unusable and only suitable for technology geeks. But it is hard to deny technology push is an invaluable innovation force too. Where would smartphones have been without multitouch technology? Do we blame the engineers developing this technology without a user need, or Apple for recognizing its potential? Just like the other forces technology push can be a major driver in new product development, but it needs to be balanced with the other forces.
By zooming out of the nasty details of integration in specific projects, the integrative innovation model is a design management tool, more than a design tool. It helps planning design projects by focusing on the most important viewpoints and contexts. It can be a useful checklist in several ways. It shows which skills you need in the team, it allows for a constant ‘integration’ check when the project is running and it guides the research questions of the projects. The model has disadvantages too. User, technology and design may work for interactive product development, but often business is just as important. Also the set of forces and contexts may be quite different for other disciplines such as media design or social design. Also, the model doesn’t solve the difficulty of integration in design. In most of the projects I did, new viewpoints arose along the way. These could not always be ‘fitted’ on the original big three of the integrative innovation model, but where just as important for preserving an integrative solution. But then again, once we found those “project specific viewpoints”, we were always close to the finish, and design management was no longer a top priority. Until that point the model served us well.
I wrote about the integrative innovation model before in our The Web And Beyond paper: “UX in the Wild: on Experience Blend & Embedded Media Design”. In this paper you can find examples of projects for which the model served as a starting point.
The paper by Charles Houde & Stephany Hill is available online as well.
Filed under: pointer, design education design | Leave a Comment
Tags: Innovation, integration, technology push, design push, user pull, user centred design, design management, Charles Houde, Stephany Hill
I do not tend to think about netiquette a lot. As with privacy, I always feel that etiquette on the web should be debated by a bunch of gray conservatives; or at least by people who are much wiser and more decent than I am. But, when I finally started thinking about it, I figured the wise bunch were missing a vital point. So I decided to help them out. The point is this: language use, even on the web, is a form of coordinated action. In this post I will explain what I mean by coordinated action, why this view on language use is so important and how it affects netiquette. I will argue that language use works if both sender and recipient share a responsibility, and why, in the daily practice of internet use, I feel it is the recipients who need to step up – and be more, well, decent.
Decent, however is not a word you would use for sender which planted the seeds for this post. This boy, a student, said something really nasty about one of my colleagues on Facebook. I never learned his exact words, so for the sake of argument: let’s assume it was beyond all limits, it was not something he would say to a teacher in real life and certainly not something he should have said at all. Most Facebook post go by fairly unnoticed, but attention sticks to the outliers. This particular post was picked up by a social media scanner that my institute uses to oversee what people say about us online. Next, because of the aggressive tone of the message it was forwarded to the director of my department. He felt the incident was serious enough to write all students (and staff) an e-mail, alerting them to basic netiquette rules and the social media code of conduct of the institute. This is how I learned about what happened. Now, I tend to be quite liberal about what you can say online, but in this case it was easy to put myself in my colleagues feet as well – this aggression could just as well been directed to me. So it took me a long time to figure out whether I had more sympathy with the student in question or with my director pressing students in general to be more sensitive about the type of information they put ‘on a public website’. In the end I decided the student deserved my sympathy most.
To me, more than a generation gap between the n-geners and the older generation this little Facebook incident is an example of a widespread confusion about the way human communication works. Most people see communication as a transfer of information. There is a sender who encodes his communicative intention into a message which is transferred through a channel and decoded by a receiver. If something goes wrong it is because of noise, which could be an encoding or decoding problem – or it could be a property of the channel. You can look at communication this way, but there is an alternative view which is often much more suitable: Herbert Clark’s theory of language use as joint action.
In a nutshell, Herbert Clark’s theory claims that using language is like dancing the Tango. For it to work, both partners have to play a part, in close coordination with each other. Dancers place their bodies and feet in response to their partners moves, to move the dance forward and similarly speakers and listeners work closely together to advance their conversations. To see how, it could be useful to discuss the actionladders for speakers and listeners briefly.
Produce sounds Attend to sounds
Give a signal Recognize signal
These action ladders are a layered protocol for human communication, but what is special about Clarks view is that he shows speakers and listeners have to cooperate and coordinate on all levels of the action ladder simultaneously. If I propose to my wife “Let’s go for a walk” we need to coordinate on four levels. I need to get her to attend to the sounds I am making (1), make her recognize that I am trying to tell her something (2), pick words that she can understand (3) and put it into words which make her consider my proposal (4). At all four levels we both need to play our part and monitor each other to see if we are still on track. We do it effortlessly every day, but using language is a very intricate and coöperative performance.
The model also applies to written communication. Clark sees face-to-face communication as the basic form, but he believes mail, chat and traffic signs are forms of joint action too. When I try to write a letter to a distant girlfriend, I weigh my words so she will understand them rightly. Often I wished she was there to give me direct feedback. Online chat often goes awkward, because temporality is messed up, but we can set things straight easily. WIthin our country we agree in the meaning of traffic signs, and teach them to all drivers, so they can communicate without words when they are on the road. In case of an accident any misunderstanding of the socially agreed meaning of these signs is taken very seriously. All these forms show human coordination, although it may be indirect and the forms may be more static than in everyday conversation.
All this coordination would not be possible without a lot of background knowledge, which Clark calls “common ground”. Possibly the most important contribution of his theory to our understanding of human communication is that he shows how we build up this common ground over time by talking to each other. For example, if my mother tells me: “the doctor thinks I am in good shape”, I know she is talking about the professional opinion of her physician and not her sexual appeal for to doctor across the street. This has to do with my background: I didn’t misinterpret her words because I knew she went for a checkup. But my mother knew, I knew about her doctors visit, because see told me about it yesterday. This is language coordination at work, before my mother picked her words, she considered what my background would be and she did not have to be afraid for any misunderstandings.
What does this mean for netiquette and the role of the recipient? The difference between an information centric model – the sender, receiver noise model – and a coordination centric model, such as Clark’s common ground theory, becomes obvious when a post is read by someone who is not part of the originally intended audience. The director of my institute in the example I gave earlier for example.
In the traditional information centric view recipients just need to decode messages they have access to. This means, that if you put something on a public website, whole world is entitled to read, interpret and have an opinion about it. Access defines the audience: it is the sender who takes all responsibility, including managing access, the receiver has a free lunch. In a coordination centric view, we admit that speakers or senders are trying to coordinate meaning and understanding with a specific group. The language in this group is defined by the common ground they created, thus by the conversational history they share. It is no longer enough for the receiver to have access to the information of the sender: if he is not part of the intended audience, he does not have the conversation history, so he cannot understand and appreciate the true meaning of the post. This doesn’t have to mean, listening in is always inappropriate, bad or evil, but it seems reasonable to ask the uninvited receiver to suspend his judgment until he has verified the true intentions of the speaker.
When the n-gener considers his Facebook page to be his own private channel on which he can say whatever he likes, he is not making a controversial cultural statement or showing a bad understanding of this technology or his privacy settings. No, he is stating an empirical fact: for all post he has written so far, he only has gotten responses from an ingroup of friends. It is reasonable to assume they will be the only ones to respond when he is angry and decides to put it in words he should not have used in the first place. In turn, my colleagues (including the director of my institute) and me are justified if they believe these words are inappropriate considering our own background knowledge. But if we want to do something about it, we need to invest in normal online human relationships with these students: we need to make proper friends or follow them, we need to make ourselves known, we need to engage regularly, not only if we dislike what is being said, so that we build up a discourse and earned respect if we disapprove of something. If we learn about inappropriate posts using a scanner, and start picking on decontextualized outliers, it would be decent to be extremely modest. Let try private messages of the sort: “sorry for spying on you, my scanner told me you said such and such, what did you really mean by that comment?”
I have written about a coordination view on language use in my last post: A Case for Privacy Coordination. Also I briefly discussed Herbert Clark’s theories about language as coordinated action in my post: Does Twitter have a Tempo?. I wrote about n-geners in my post Evaluating the Net-Gen Argument.
Of course Herbert Clark’s comprehensive book “Using Language” provides the ultimate background to these ideas.
Filed under: (re)thinking media, probe | 3 Comments
Tags: Action Ladders, Communication, Herbert Clark, Language Coordination, N-geners, Netiquette, Recipient Responsibility, Using Language
Something fairly insignificant happened on the train the other day, when it occurred to me there is something wrong with the way we think about (and design for) privacy. It wasn’t long after the introduction of the “OV-chipkaart”: an electronic payment system for public transport in the Netherlands. I was minding my own business, when the train attendant asked a fellow traveler whether she had been on the Schiphol airport two days earlier. She replied that she was; the train attendant said that something had gone wrong with check-in there; my fellow traveler shrugged and this was the end of the conversation. Nothing more happened, but I was stunned. I do not consider public transport travel history as superprivate information but this short exchange showed three privacy flaws in the card and the way it is used. First, travel information (up to 12 trips) is stored on your travel card for no obvious reason, second, train attendees see this information whenever they scan your card and third, the train attendee in question did not feel inhibited at all to share information this publicly; or at least with travelers nearby. And it is not an exceptional incident. Privacy flaws like these three are becoming more and more common while it is possible to prevent them. In this post I would like to explain how.
I had never considered privacy as something important you could (or should) design for, but I changed my mind because of two PhD theses from co-workers. At 22 September 2008, Natalia Romero Herrera presented her thesis titled: “Coördination of Interpersonal Privacy in Mediated Communication” at the Eindhoven University of Technology. Less than a year later, May 19th, 2009, her co-worker Evelien van de Garde-Perik followed with her thesis: “Ambient Intelligence & Personalization: peoples perspectives on information privacy”. Both these theses are worth reading, but they are for a specialist audience. Still, I believe there is one bit of knowledge from this work that every designer needs to know about. This is wat it is: every designer should be able to make the distinction between privacy coordination and information privacy. I believe the difference between these two is easy to grasp, but it matters a lot which one you choose to frame the privacy problem: depending on your choice you will ask different design questions.
Let me take a closer look to the distinction between privacy coordination and coordination privacy. Most people think of privacy as an information privacy problem. Van de Garde defines it as ‘the ability of the individual to control the terms under which his or her personal information is acquired and used by others’. If you want to control your information privacy, you need to decide who can have and use information about you. Van de Garde dates the privacy debates back to the 19th century, when recording equipment such as print and photography emerged. These new technologies made it possible to present information about someone to a new audience after the event was over. To my taste it is this possibility of second hand (mis)use which gives information privacy its ungraspable and ghostly feel. Debates on information privacy are often about situations that do not really happen, or at least not often enough to assess the risks reliably.
This is also why it is difficult to design for information privacy but Van de Garde’s research shows it is important. She shows people do have privacy concerns and show a limited understanding of information about privacy consequences. Also it is difficult to give users control in a manageable way. People would like to know: what information about them is collected or stored (1), who can be the possible recipients of their information (2) and the purpose of use of their information (3). If you want to give users control over all these properties, for several bits of information and multiple possible audiences you end with a combinatorial nightmare. So, Van de Garde evaluated several interaction models which could simplify this task. Users reacted diversely but there are clusters of with similar privacy preferences visible in the data. Van de Garde did not give a full classification of users based on their privacy preferences. Such a classification would be interesting though. Designers privacy sensitive interfaces could create personas out of them. Designing for information privacy is a necessary but difficult task because, also in Evelien van de Garde’s work privacy concerns turn out to be so context specific. It is this finding where design for privacy coordination enters the picture.
Surely the privacy discussion had evolved since the 19th century, but in 1975 a social psychologist called Irwin Altman gave it a radically different twist. According to Altman privacy is “an interpersonal boundary process by which a person or a group regulates interaction with others”. Altman argues that, to understand privacy, we must see it as an ongoing coordination process between people rather than an information problem. In an everyday privacy coordination problem in the office you need to cooperate to set privacy boundaries. You may open your office door when you’re open to talk to colleagues and close it when you are not. You colleagues may barge in, despite the closed door. You may ask them to leave if they do so; and so on. Or, to give a different example: you didn’t really want to tell about an embarrassing situation on that party, but you didn’t want to spoil truth or dare, so you told it anyway. Rather than talking about what might happen to your data in the future, privacy coordination is about telling people to keep distance or come closer here and now, which we do every day – in a myriad of way’s. When we look at privacy this way – as an everyday coordination problem, information privacy is a special case where we try to capture and arrange many of these problems in a couple of fixed rules. Information privacy stands to privacy coordination as law stands to the resolution of everyday conflicts; it is its sediment.
Thus, Altman’s views bring privacy closer to home, but are they applicable to design? This is the question which Natalia Romero Herrera explored in her thesis. Apart from Altman, she build on Herbert Clark’s common ground theory and she did extensive fieldwork, both to create a model to describe and understand how people can communicate privacy borders in mediated communication: the privacy grounding model. Most electronic tools are ‘always on’ technologies, and do not offer privacy grounding possibilities, such as the ‘doorpolicy’ which I just discussed. You can close your e-mail client, but there are few tools which allow you to postpone unimportant emails or mails from certain people. Imagine an email clients which allows potential senders to know the settings of your privacy filter and to breach it when it is important enough. In everyday conversation we solve this type of coordination problems almost effortless, and non-verbally. The design difficulty for electronic privacy coordination support is to keep it as lightweight. It needs to be in the background, it needs to be easy to control and it needs to allow for ambiguity. Using the privacy grounding model, Romero did built tools which support lightweight communication of privacy needs. She thought of a single click and a drag and drop privacy coordination solution for instant messaging applications and she designed a tangible “availability cylinder” to express your general availability level in an easy way. So, Romero did find ways to design for privacy coordination.
So let me go back to the train and the “OV-chipkaart”. There is information privacy design in the card system. Someone has decided that all cards would store 12 trips and these would be available to all cardreaders . The difficulty of this design decision is that it has to cover all possible usecases of the card and their social contexts. However, what van de Garde and Romero show, is that after the decision about information storage has been made, there is ample room for privacy friendly user experience design. Evelien van de Garde’s, work shows that it would have been possible to give users control over the situations and people with which they were willing to share this information. Natalia Romero’s work suggests to design the interfaces so people can coordinate whether they want to share their information when they need to. Say the travel history would have been on a second screen of the train attendants interface. Let’s also assume the railways would have created a protocol urging train attendants to treat this information with care. This way exchange in my train that day could have been much more pleasant. Attendant: “I see there is something wrong with a previous check-in, would you mind if I see what went wrong”. Traveler: “sure, I would like to know”. Attendant: “There is a failed check-in at Schiphol, probably you flipped your card to fast past the reader”. Traveler: “Al right good to know”. There is the same information on the card in the example. The difference is that users can coordinate their privacy needs directly when they need to because this process is supported in the interface. Thinking of privacy design as a privacy coordination problem puts privacy concerns in the actual context of use, in the here and now and in the hands of the users. If applied with care this makes the life of the designer easier and the life of the users a lot more private.
I haven’t written extensively about privacy before, although my post about Diaspora, the open source alternative for Facebook, touches on the issue and provides some links. Also I briefly discussed Herbert Clark’s theories about language as coordinated action in my post: Does Twitter have a Tempo?
The full text of both Coordination of Interpersonal Privacy in Mediated Communication and Ambient Intelligence and personalization: peoples perspectives on information privacy can be found online.
 Cited from Nathalia Romero’s thesis pp28
Filed under: (re)thinking media, discussion | 3 Comments
Tags: eindhoven university of technology, Evelien van de Garde-Perik, Herbert Clark, Information Privacy, Irwin Altman, Natalia Romero Herrera, OV-Chipkaart, phd theses, Privacy, Privacy Coordination, User Experience