A definition of technology s-curves drawn from Clayton Christensen (in this paper):

“The technology S-curve has become a centerpiece in thinking about technology strategy. It represents an inductively derived theory of the potential for technological improvement, which suggests that the magnitude of improvement in the performance of a product or process occurring in a given period of time or resulting from a given amount of engineering effort differs as technologies become more mature.
(…)
It states that in a technology’s early stages, the rate of progress in performance is relatively slow. As the technology becomes better understood, controlled, and diffused, the rate of technological improvement increases . But the theory posits that in its mature stages, the technology will asymptotically approach a natural or physical limit, which requires that ever greater periods of time or inputs of engineering effort be expended to achieve increments of performance improvement. “

Why do I blog this? Given that I use this tool more and more often in talks, workshops and work, it’s good to get back to the literature and understand it more thoroughly. In some work recently I mostly used it to describe evolution of certain technologies such as location-aware systems, 3D virtual worlds or mobile gaming. Generally, the point of is to describe a succession of waves starting from an idea as shown on the picture below. For instance, with the “location-awareness” idea, the first wave of mature products was navigation systems (quite often found in cars with garmin and tomtom devices), a second wave concerns place-based annotations systems or people finder (in that case, nothing’s really mature in the same sense as the first wave). Besides, I am well aware of the limits of such curves but they offer a relevant way to discussion diffusion of innovation.

Reading “Manuel de prospective stratégique, tome 1 : Une indiscipline intellectuelle” (Michel Godet), there was an interesting chapter about the most frequent error when doing foresight.

General causes are:
1) Forgetting change (over-estimation) and inertia (under-estimation).
2) “Announcement effect”: some predictions only aim at influence the evolution of the phenomenon and then contribute to its realization
3) Too much information (noise), few strategic information
4) Inaccuracy of data and instability of models (one should always ask whether a small modification in input data will change the output)
5) Error of intrepretation
6) Epistemological obstacles (looking at the tip of the iceberg / or where the light is)

Specific causes:
1) Uncomplete vision (leave behind other variables, disruptions, new trends…)
2) Excluding qualitative variables (that cannot be quantified
3) Thinking variables have static relationships
4) Explaining everything by looking at the past
5) Single future
6) Excessive use of mathematical models (mathematical charlatanry)
7) Conformism to gurus

To complete my notes on the LIFT07 workshop about foresight, there is a very dense and insightful podcast of Bill Cockayne’s talk at Design2.0 (mp3, 15.41 Mb).

In this talk, Bill explains that one of the challenge for take companies/student in engineering schools is to get people understand the bigger context, complexity and big systems. Bill started as a technologist and migrate as a technology-foresisght/strategy person. His point is to ask questions such as “where is it going?“, “why is it going there?“. This is not a matter of being a futurist, not about predicting anything but rather to work on “how do you think about this coming technology?” “how do you think about this coming social change?”. Technology sometimes drive social change, does not, sometimes, maybe but the question is how do you know when?. It’s not predictions, it’s something that comes out of knowing where information comes from.

Beyond tools to design for today’s future (ethnography, brainstorming prototypes) and those for going a little further (scenario planning), the point is to go much further: how do you critically assumptions and build models. Oddly enough this stuff is simple, using 3 tools he describes. My raw notes below:

3 tools: point of view questions, X-Y graphs (out to get there by telling stories, looking for triggers of change, think about to get there when we think about what want to be, being normative (design a better future), defensive (design for a future that is coming but we don’t like it), how to prepare for that kind of things),

1) Simple rule: You won’t get there from here

let’s say you design a toothbrush, you observe current users so you’re going design today for a year from now
as you get out past 7 years that does not work, who are you going to observe?
this is all a POV: get out there with

2) X-Y graph: a structured brainstorming tool
Issues: A versus B. So discuss with your team: What would be the 2 most salient issue? issues being one on X and one on Y. So you have 4 endpoints. What would the salient issues that affect the questions we’re asking in 20 years? It is going to be perception? a social issue? no tech change? Older people (with experience) are better doing this because they’ve seen change (they felt what is 5 years).
After a whole day, you may have 10 good X/Ys. Good = something you learn over time and sth you feel intuitively.
Tell stories when you’re doing it, catchphrases, funny stories…

What you want to look for is whitespots = possibilities, you can make a difference here
either no one is going there because is difficult or it is an opportunity
how might we put something there? a toy, a computer, a social change

3) Then you start building scenarios, like design but way far out, 20 years ahead
what if have 3-5 stories? what would the world be out there?
the most important things about these lines: no changes, lots of changes, one big sweeping change…

tell a story in 5′ and then spend the rest of the afternoon going backwards,
tell me what had to happen all along the way, tell me when it had to happen, give me a timeframe, the trigger, the driver, when does something has to happen is very critical
as you being to go backward, you realize what has to happen (before a product occur, need of having another tech, so another guy has to invent this tech)

As you begin to go further out in time, you have a much harder time to say how close your change drivers are going to be. Then assume that all the decisions you make are too pessimistic and far out. In near term, assume that everything you say is too slow

Long term changes tend to have trigger than is not necessarily in the center of where the change is occurring
when economics are changing is not that because a person stood up and said “Wall Street is going into that directions” it’s more that you watch the housing data, you watch the number of kids that are how many kids are being born, breastfed,… and then you ask where is another change coming further off and how is it going to be its impacts?

The questions were quite interesting. One of the person asked what is the biggest mistake made by companies. Bill argues that most big companies forgot that research existed for two reasons: invent new things and spend a lot of money obtaining patents, the other is to have a bunch of guy who sit around, doing this kind of things he presented in the afternoon, drinking their coffee. Another issue is the fact that None of us read enough, none of us talk to smart people enough.

“Read methodologies and then read WSJ, E, NYT, CSM… daily because you needs to start getting a feel of where data comes from. You may be watching very closely where your products are going to be but something is changing in an area you never even thought but that could infect it, that could be an opportunity. READ MORE

These publications have the broadest range of readers they have op-eds. Get a broad view of business, social, economics, random technology stuff. Take the biggest daily newspaper that don’t focus on news, more like the economist, that look for the analysis, context, why this happened, why A did X… Over time you build up and ability, look for different views, it’s not a bias you’re looking for, but a a different viewpoint“

Why do I blog this? great food for thoughts, methods and ideas about how to structure what I am doing in something more formalized. Besides, the question of “data” in foresight, addressed in the talk, is of great interest to me.

Currently completing my PhD program (thesis defense is next week), it gave me the occasion of looking back and think about what interest me. My original background is cognitive sciences (with a strong emphasis on psychology, psycholinguistics and what the french calls ergonomie) and the PhD will be in computer sciences/human computer interaction. In most of my work, I have been confronted to multidisciplinary/interdisciplinarity (even in my undergraduate studies).

It took me a while to understand that my interest less laid in pure cognitive science research (for example the investigation of processes such as intersubjectivity, and its relation to technologies) but rather about the potential effects of technologies on human behavior and cognitive process. In a sense this is a more applied goal, and it led me to take into account diverse theories or methods. Of course, this is challenging since mixing oil and water is often troublesome in academia. Given that my research object is embedded in space (technology goes out of the box with ubicomp) and social (technology is deployed in multi-user applications), there was indeed a need to expand from pure cogsci methods and including methods and theories from other disciplines. The most important issues regarding my work for that matter were the never-ending qualitative versus quantitative methods confrontation (I stand in-between using a combination of both, depending on the purpose) AND the situated versus mentalist approach (to put it shortly: is cognition about mind’s representation? or is it situated in context?). So, this was a kind of struggle in my phd research.

However, things do not end here. Working in parallel of my PhD as a consultant/user experience researcher for some companies (IT, videogames), I had to keep up with some demands/expectations that are often much more applied… and bound to how this research would affect NPD/design or foresight (the sort of project I work on). Hence, there was a need to have a discourse about these 2 issues: design and foresight. No matter that I was interested in both, it was not that easy to understand how the research results/methods can be turned into material for designers or foresight scenarios. SO, three years of talking with designers, developers, organizing design/foresight workshops, conferences helped a bit but I am still not clear about it (I mean I don’t even know how to draw something on paper).

Recently, I tried to clear up my mind about this and the crux issue here is the constant shifting between research and design (or foresight, sorry for putting both in the same bag here but it applies to both). The balance between research that can be reductionist (very focused problem studied, limits in generalizing or time-consuming) and design that needs a global perspective is fundamental. The other day,I had a fruitful discussion with a friend working on consumer insight projects for a big company. Coming from a cognitive science background as this friend, I was interested in his thoughts concerning how he shifted from psychology to management of innovation/design of near-future products/strategy.

I asked him about “turning points” or moments that changed his perspective. He mentioned two highlights. The first one was the paradigm shift in cognitive science in the late 80s when the notion of distributed cognition (Dcog) appeared. Dcog basically posited that cognition was rather a systemic phenomenon that concerned individuals, objects as well as the environment and not only the individual’s brain with mental representation. To him, this is an important shift because once we accept the idea that cognition/problem solving/decisions are not an individual process, it’s easier to bring social, cultural and organizational issues to the table.

The second highlight he described me is when he use to work for a user experience company that conducted international studies, he figure out that the added value not only laid in those studies but also in the cumulative knowledge they could draw out of them: the trend that emerged, the intrinsical motivation people had for using certain technologies, the moment innovation appeared. This helped him change the way he apprehended the evolution of innovations and made him question the fact that they can follows long s-curves.

Why do I blog this? random thoughts on a rainy sunday afternoon about what I am doing. This is not very structured but I am still trying to organize my thoughts about UX/design/foresight and how I handle that. I guess this is a complex problem that can be addressed by talking with people working on design/foresight/innovation. What impresses me is observing how individual’s history helps to understand how certain elements encountered shape each others’ perspective.

The picture simply exemplify the idea that conducting design/foresight projects need a constant change of focus between micro and macro perspectives. This reflects the sort of concern I am interested in by taking into account very focused perspectives (user interface, user experience, cognitive processes) and broader issues (socio-cultural elements, organizational constraints…).

The Bell&Dourish paper I’ve blogged about last week is still sparking some interesting discussions (interestingly it’s not only ubicomp researchers but also architects).

What is interesting to me is how this discussion about focusing on the ubicomp of today and less about proximal future connects with the discussions I had with Bill after the LIFT07 foresight workshop. The “here today” versus “could be tomorrow” argument is indeed one of the underlying questions of foresight versus design practice. In Bell and Dourish article, the authors critique these earlier visions of a proximal future not to complain about past visions, nor to understand why we haven’t gotten there but rather because it allows them to question an important assumption made by ubicomp researchers: the coming of a so-called seamless world with no bugs and perfect could of connectivity (that do not hold true as Fabien described it at LIFT07).

So the point here is the importance of the “why question”, the crux issue that the LIFT07 workshop addresses; critical foresight is about asking why something worked, why someone would want the future you propose or why the path proposed is possible. In the context of this ubicomp paper, some additional questions about the future of ubiquitous computing can be asked: what would we want: a short term vision of the next incrememental ubicomp ‘project’ or a new strong vision (as Weiser’s calm computing was). But what might be needed for having this strong vision is clear and lucid description of the why that eventually lead to a point people could aim at.

So there could be an interesting exercise to think about when criticizing the intelligent fridge, CAVES, intelligent assistants or other ubicomp dreams that failed. That could be a good agenda for a possible workshop at some point.

Some notes about the foresight methodologies discussed at the LIFT07 workshop “Re-designing the city of the future” that I co-organized with Bill Cockayne last week.

The purpose of the workshop was to a gather an heterogeneous crowd of people to discuss topics regarding the city of the future. The point in preparing this workshop was also to deal with new methodologies, to better structure foresight ideas (for instance to go beyond the design scenarios developed in the past series of blogject workshops). This is why I teamed up with Bill who gave an insightful presentation of critical foresight tools that I describe hereafter.

As opposed to design (i.e. build/invent/create), foresight is about critically explore assumptions, build models & develop questions about the long term future. One of the pre-requisite of the workshop was to read various papers that exemplified different visions of the future: Fast, Huge and Out of Control, Metropolis (1999), ‘Future Cities‘, Time, 1929 and ‘January 3000 A.D.‘, Harper’s New Weekly Bazaar, 1856. The reading of those papers was meant to spark some discussion about critical foresight: Did any of the authors did each guess correctly? If no, why were some guesses so bad? What were the changes? Were these changes social or technical? Was it a driver or a reaction? Global or local?

Then Bill introduced the first “tool” in the form of a petal graph. This is basically a diagram where he mapped each critical aspect of change that the group listed. The goals is then to find the commonalities of all these aspect, what goes in the center of the flower.

The petal graph is indeed a good tool to realize how the future is a complex problem. One one hand, it’s uncertain (not measurable). On the other hand, it’s ambiguous and we do not even know what to measure. However, this does not mean that we cannot make assumptions: “You can’t predict the future, but you can invent it” as the motto says. The point here is not too do futurism but to look at data and use analytical reasoning to discern what might exist and what we could build. Thus, the value do not lay in predictions but in the underlying discussions: the “why” of predictions: we thus focus on the questions generated, not the answers. This said, the crux issue in foresight is to be critical about what others says about “the future”. This is why we looked at different material, be it press article, journal papers or the Walt Disney’s EPCOT center video. In a sense, the main goal is to explore, deconstruct, and critique the futures envisioned by others as a methodology of understanding, using a multidisciplinary approach. The following step was to use three tools for foresight thinkings: S-curves, x/y axes and white/hot spots.

The s-shaped curve is the canonical representation of how an invention evolves over time from the idea to the mass-market commercialization (plateau) with every technologies/instance that occurred in between (and caused the raised of the curve). This tool enables the discussion about the social/technical changes that allowed this progression. Lots of questions can be asked using this curve: Why do so few futures seem to follow the path? This helps contextualizing what’s going next.

Then we picked up 2 dimensions/topics that can interact and represented them on cartesian axes.

The choice of these axes is important since it is meant to generate “questions”. Once the axes are defined, this is a tool to discuss stories/concepts/inventions and position them in the quadrants according to the 4 dimensions that has been set. This allows to have white spots that can be considered as opportunities (or they don’t exist for a certain reason that should be discussed) and hot spots with a high density of existing examples.

Based on white/hot spots and depending on the time range, one can then unfold the history backward as represented on that picture to answer the question: how did we get to this spot, when were the changes? Doing so need to think about early indicators of change, whether those changes are already in view, what type of events? where would this events be likely to occur.

Once this was done, Bill introduced tools for “Foresight Thinking for Designing”: observe, analyze and prototype. Observing is a matter of thinking about people today and at future time: assuming that people will change, what would be the reasons/motivation/driver, when thinking about change what are the early indicators (triggers or incipient)?. The analysis part is mostly about questions: ideas are fine but questions are more important and assumptions critical. Finally, the prototyping part concerns the models but also the underlying assumptions, questions, and changes. The best models generate questions around the areas of highest change. And finally the last step is to communicate, which can take various forms: stories (short stories, speculative fiction, science fiction, counterfactuals), scenarios/personas, movies, maps (Cross-impact, Trends, S-curves) or even tangible artifacts.

We then constituted 5 groups who had to use the previous tools to had to develop a future to report about the “city of the future” and tell this to the others at the end of the workshop. If I have time I’ll post about the workshop results but to me the most important thing was the discussion it fostered (especially among groups).

Finally, after a LIFT I managed to have more time for reading good papers such as Yesterday’s tomorrows: notes on ubiquitous computing’s dominant vision by Genevieve Bell and Paul Dourish (Personal and Ubiquitous Computing, 2006).

The paper deeply discusses Mark Weiser’s vision of ubiquitous computing, especially with regards to how it has been envisioned 10 years ago and the current discourse about it. In fine, they criticize the persistence of Weiser’s vision (and wording!). To do so, they describe two cases of possible ubicomp alternative already in place: Singapore (example of a collective uses, computational devices and sensors) and South Korea (infrastructural ubiquity, public/private partnerships).

Their discussion revolves around two issues. On one hand, the ubicomp literature keeps placing its achievements out of reach by framing them in a “proximal future” and not by looking at what is happening around the corner. Such proximal future would eventually (for lots of ubicomp researchers but also journalists and writers) lead to a “seamlessly interconnected world”. The authors then express the possibility that this could never happen (”the proximate future is a future infinitely postponed”) OR more interestingly that ubiquitous computing already comes to pass but in a different form

On the other hand, ubicomp research is very often about the implementation of applications/services, assuming that the inherent problems would vanish (think about privacy!).

Therefore, what they suggest to the research community is to stop talking about the “ubiquitous computing of tomorrow” but rather at the “ubiquitous computing of the present”: “Having now entered the twenty-first century that means that what we should perhaps attend to is ‘‘the computer of now.’’“. Doing so, they advocate for getting out of the lab and looking at “at ubiquitous computing as it is currently developing rather than it might be imagined to look in the future“. And of course, they then points to an alternate vision that Fabien discussed last week at LIFT07:

the real world of ubiquitous computing, then, is that we will always be assembling heterogeneous technologies to achieve individual and collective effects. (…) Our suggestion that ubiquitous computing is already here, in the form of densely available computational and communication resources, is sometimes met with an objection that these technologies remain less than ubiquitous in the sense that Weiser suggested. (…) But postulating a seamless infrastructure is a strategy whereby the messy present can be ignored, although infrastructure is always unevenly distributed, always messy. An indefinitely postponed ubicomp future is one that need never take account of this complexity.

So what’s the agenda? Based on William Gibson famous quote about the future being there and not evenly distributed, they encourage that:

If ubiquitous computing is already here, then we need to pay considerably more attention to just what
it is being used to do and its effects.
(…)
by surprising appropriations of technology for purposes never imagined by their inventors and often radically opposed to them; by widely different social, cultural and legislative interpretations of the goals of technology; by flex, slop, and play. We do not take this to be a depressing conclusion. Instead, we take the fact that we already live in a world
of ubiquitous computing to be a rather wonderful thing. The challenge, now, is to understand it.

Why do I blog this? Best paper for weeks. This particularly resonates to the way I think about Ubicomp… meaning that no the recurrent intelligent fridge some have dreamed of 10 years ago is not the “fin de l’Histoire” (end of History). I really like when Bell and Dourish bring forward issues like ubicomp can rather be exemplified as Cairo with its freshly deployed WiFi network set to connect all the local mosques and create a single city-wide call to prayer than having a buddy-finder locator.

Moreover, the authors express their surprise to the fact that researchers are still positing much the same vision as years ago. This reminds me the ever-decreasing time-frame futurists tried to predict: the year 2000 was really the ending point and prediction were always targeted to that period. Now that we’re in the (so-called?) 21st century, it’s as if there could be no other future.

Anyway, that’s a call to go “on the field” and see what’s happening and the effects of technologies.

Today starts the LIFT07 conference with a day devoted to workshops. Along with Bill Cockayne (Stanford Center for Critical Foresight), we co-organized a session entitled “Re-Designing the City of the Future”.

This workshop will begin in the use of existing futures and critical foresight methods to understand how the future is being envisioned. Then using design research tools, participants will design social, technical, and business innovations that could exist in 2015 and 2025 regarding the future of the City. Key skills will be the integration of analysis with experience, foresight with design thinking, and building and communicating prototypes for the future.

More about that later.

In the last issue of Strategy Business, there is a long in-depth interview of Alvin Toffler about his the book he wrote with his wife “Revolutionary Wealth”. While the whole interview is a must-read for people interested in foresight, innovation, prosumer revolution and stuff like that, I was bemused by this excerpt:

S+B: When I looked at Future Shock recently, I was surprised at your stridence. You wrote of the acceleration of the pace of change as an illness, “a cancer in history.” With 35 years of hindsight, would you still describe our situation that way?

TOFFLER: Well, I might tone down some of the language. I was 35 years younger. But I think the basic argument of the book stands. We’re always asked what we got wrong, and we did get a few things wrong. That’s inevitable when you’re looking 30 or so years ahead. The hardest thing to forecast is timing — when certain events would happen. We said, back then in 1970, that humanity would clone animals, and that has happened; we said that we would also clone humans, and I still think that’s likely. But we were wrong in the timing. We said that these would happen by 1985. We didn’t make that date up. We got it from one of the world’s leading Nobel Prize–winning biologists, who happened to be rather more optimistic than he should have been.

There’s another passage in the book where we talk about throwaway products, that someday we may be wearing paper clothing. And we aren’t. Yet.

I always get a laugh from an audience when I say, of course, we futurists have a magic button; we follow every statement about a failed forecast with “yet.”

Why do I blog this? the power of words always strikes me.

A quote I like, taken from “What If Our World Is Their Heaven? The Final Conversations Of Philip K. Dick” (Gwen Lee, Doris Elaine Sauter, Tim Powers). Also pointed by Tim Powers in the introduction of the book:

“That’s another technical advice, you casually have one character say to the other, ‘Where did you put the biochips?’ ‘I put them back in the cupboard where they belong.’ That’s all you need to say… See it’s amazing how easy it is to write if you know how…

Why do I blog this? PKD’s was of describing the future is awesome, I really like the way he only drop words and concepts that build a sort-of ambiance of the future.

Seen last month in CNN Money, this article describes how the through the combination of satellite maps and 3-D software (the 3D modeling program SketchUp), Google Earth is turning into a virtual online playground. Some excerpts I found interesting below. It starts like the Second Life crazyness:

You can already download user-generated layers that sit on top of Google’s 3-D Earth and show you, for example, the location of celebrity houses or hiking trails or famous landmarks. One dating service has even started showing people looking for partners as a Google Earth layer. Real estate companies have started showing off virtual versions of their buildings (for sale in the real world) on Google Earth. SketchUp allows them to build entire models of their apartments, right down to the microwave oven.

The result could be that we’ll soon populate a virtual version of planet Earth instead of the made-from-scratch metaverses like online games or Second Life. The main element Google Earth is missing today is avatars (…) “I would expect to see someone using Google Earth as a virtual social space by the end of the year,” says Jerry Paffendorf, research director of the Acceleration Studies Foundation

Then the article starts describing how the Web can become a 3D metaverse-like environment with blabla and stuff that I am still dubious about.
Why do I b log this? Even though I am not very enthusiastic about the whole article there are some relevant stuff here. Of course, stories such “Consumers could fly into the virtual New York, go shopping in a virtual Times Square, get past the velvet rope at a virtual Studio 54 and chat with an avatar dressed as Andy Warhol” always get my hackles up. The journalist seems to stretch out a bit his conclusions. IMO what is interesting with google earth and sketchup is the creativity it allows not that it can be the basis for the future of the web. This said, I additionally think it’s very interesting to have a non-avatar based virtual environment; it’s indeed a model on which interesting things could be done (though I feel like some avatar will pop up at some point).

A quote from Posted in Future | No Comments »