Tangible Models

MIT researchers have developed a cool digital drawing board that simulates the physics of simple systems:

You can play with something like this with Crayon Physics or Magic Pen. Digital physics works because the laws involved are fairly simple, though the math behind one of these simulations might appear daunting. More importantly, they are well understood and universally agreed upon (except perhaps among perpetual motion advocates).

I’d like to have the equivalent of the digital drawing board for the public policy and business strategy space: a fluid, intuitive tool that translates assumptions into realistic consequences. The challenge is that there is no general agreement on the rules by which organizations and societies work. Frequently there is not even a clear problem statement and common definition of important variables.

However, in most domains, it is possible to identify and simulate the “physics” of a social system in a useful way. The task is particularly straightforward in cases where the social system is managing an underlying physical system that obeys predictable laws (e.g., if there’s no soup on the shelf, you can’t sell any soup). Jim Hines and MIT Media Lab researchers translated that opportunity into a digital whiteboard for supply chains, using a TUI (tangible user interface). Here’s a demonstration:

There are actually two innovations here. First, the structure of a supply chain has been reduced to a set of abstractions (inventories, connections via shipment and order flows, etc.) that make it possible to assemble one tinker-toy style using simple objects on the board. These abstractions eliminate some of the grunt work of specifying the structure of a system, enabling what Jim calls “modeling at conversation speed”. Second, assumptions, like the target stock or inventory coverage at a node in the supply chain, are tied to controls (wheels) that allow the user to vary them and see the consequences in real time (as with Vensim’s Synthesim). Getting the simulation off a single computer screen and into a tangible work environment opens up great opportunities for collaborative exploration and design of systems. Cool.

Next step: create tangible, shareable, fast tools for uncertain dynamic tasks like managing the social security trust fund or climate policy.

5 thoughts on “Tangible Models”

  1. Pretty close. The “wheel” is actually a puck that you plop down on any variable or constant in the model. And, there are lots of pucks. When you put any of the pucks on a variable (or constant), the puck becomes the control for the variable (or constant) — in a sense the puck “becomes” the variable (or constant). Then, you can gesture with the puck to do a number of things — you can rotate the puck, which changes the value of the constant (or even a variable) just like synthesim (this design actually predated the introduction of synthesim in Vensim). A different gesture will give you a list of alternative molecules that you can substitute for that variable or constant — e.g. you can substitute an aging chain for a single stock. You can also see the molecules of which the variable (or constant) is a part within the given model (a single stock might be part of an aging chain, and also part of a co-flow in the same model), choose one of those molecules and then see alternative molecules that you can substitute — for example, an alternative molecule for an aging chain might be a molecule that also includes the ordering function for the supply chain.

    The “many” pucks feature is actually cool. When you use both hands, each on a puck, it feels like your reaching right into the model.

    The innovations are:

    1) The TUI (which consists of pucks, a “magic table”, and a projector mounted from the ceiling that projects onto the table).
    2) A replacement hierarchy of molecules. This is what lets you replace one molecule for another (and remember a stock is a molecule and so is a full model) and always end up with a working and semantically logical model.
    3) A data base of molecules (arranged in that hierarchy) and associated machinery that allows you to easily “find” the molecule your looking for and which automatically adds every new structures (i.e. new molecule) you create via substitution. So the data base of molecules is constantly getting richer. This machinery automatically “cannibalizes” the model your building for parts as you build it — and the parts are then available when you next want to model something.

    Obviously, the TUI is just one possible interface onto the hierarchy and database (a very cool one); we also had an interface that was more normal (i.e. entirely CRT and mouse based).

    People from four MIT groups and Intel collaborated in the project: The Center for Coordination Science (Tom Malone, George Herman, and John Quimby), the Media Lab (James Patten and Hiroshi Ishii), CTL (Jim Rice), Intel (Mary Murphy-Hoye), and the System Dynamics Group (Paulo Goncalves and Jim Hines).


  2. Hi Jim,

    It is always the serendipities that generate new thinking. This time it was a visit to my former professor in logistic (he would be surely interested in what you have achieved) while searching for the “Systems Thinking Playbook” we came across Dennis Meadow’s Sustainability Institute. This link I visited today at home, found Drew’s blog and while reading the story on the “3 Climate Negotiators” I ran across Tom’s blog and now found this great post.

    Actually that kind of tangible models I would like to use in the forthcoming teaching Lean Thinking to people. Combined with system dynamics this really looks VERY INTERESTING.

    Looking forward to hear more on the course at WPI aroud this project or directly.

    Best regards


  3. Two years later this blog entry has caught my eye again!

    Currently working on http://bit.ly/WikiWallWhitePaper (http://mindbroker.de/wiki/WikiWall). The intention is to bring people’s intangible knowledge into tangible connections, and actions alike.

    Collaborative knowledge through using multitouch, and swarm intelligence tools to bring the knowledge to use. This is especially relevant in solving the tough challenges that face humanity in the coming years, where information is often held by individuals and finds no easy way into the larger community. Nowadays the tools and methods enable new kinds of knowledge sharing, which has to be learned first (all ages and kinds of people will be engaged in this process, as knowledge is embedded in each of us, which is unique in a sense).

    Wishing everybody who are reading this a Happy New Year and a good start in 2011!!!

    1. Cool idea. Reminds me a bit of a combination of Freebase and Facebook. I especially like the idea of a museum wikiwall – having been to too many big institutions where information is fragmented onto a million tiny wall plaques, with no overarching structure at all.

Leave a Reply

Your email address will not be published. Required fields are marked *

− 2 = 4

This site uses Akismet to reduce spam. Learn how your comment data is processed.