Skillfully combining two blog posts in one I figured I’d post about the submission I put together for the Google 10^100 contest [ http://www.project10tothe100.com/ ]. Hopefully that isn’t too self laudatory. But I strongly encourage everybody else to do so as well - perhaps we can get one of our proposals into the top 100 at least! As well it’s a good exercise in formalizing ones thoughts.
One of the interesting possibilities about the digital era we live in is that we can start to bring computation to bear on every day problems. There’s probably a good chance that we’re going to see some real changes in how we make decisions in our communities. For example voting is likely to become more granular and more common - as the cost of voting becomes lower. We’re likely to have better ways of sharing knowledge - even more so than we have now - as filtering improves. And in particular Systems Science, such as the work being done at PSU, promises to give us some tools for understanding the complex interactions between land use policy, environment, economy and social welfare.
The proposal I submitted was very much in line with this last point. Here it is with parody inserted:
What one sentence best describes your idea?
(At this point I was thinking - wait a second - who is saying that? This is kind of freaking me out. Is that Larry Brilliant? Who wants to know? )
Um, ok, to create digitally augmented decision making tools for citizens by combining ideas from SimCity and Google Earth.
Describe your idea in more depth.
(I figured I could get into this and the boldface font is pretty cool. )
The core of the work is to build a web appliance that computationally simulates forward in time simple models of watersheds and human activity based on human land use. Outcomes can’t be coerced, they simply fall out of the initial conditions.
These kinds of models do exist in the Earth Sciences and System Sciences academia and the game “SimCity” as noted is pertinent. The real difference is that this would be free, public and durable.
Typical usage is that a participant would sign on, and either start a new simulation or clone an existing one, or join in to collaboratively edit a shared one. Participants would have access to available public map data but could also visually drag and drop in rivers, streams, estuaries, green space, urban space. Specific animal habitat, observed salmon populations, known bird sanctuaries and the like could be specified. Zoning such as urban, or farming, organic or industrial could be specified. Superfund sites could be indicated.
Participants could wire up relationships between things such as pesticide / sediment run off into streams, stream clarity or blockages and salmon populations, industry and carbon emissions, even speculative relationships such as between crime and green space, or airplane noise impact on animals, traffic between work and suburb, relationships between available bicycle routes and car traffic.
Then outcomes could be tested. The simulation could be fast-forwarded to a specific time. Questions could be answered according to various metrics. Is the cost of a new civic park worth it? Is there an economic benefit to a Liquid Natural Gas Plant? Is a field better used for light industry or a park or suburban housing? Are urban boundaries effective? Should Light Rail Transit be introduced? Is traffic going to be congested?
What problem or issue does your idea address?
(Problem, not problems? Hmmm, uh, wasn’t it obvious, can’t anybody READ!? Ok, clearly they want more - can’t they just google the issue? )
This tool initially would be used to model very specific problems between a set of stakeholders - and later it would become a tool that could be used by entire communities to model entire watersheds or regions.
In the former it could be used to share and examine very specific issues. For example Astoria, Oregon has had an ongoing resistance to a liquefied natural gas import terminal on the Columbia River. There are significant risks, but as well of course it does promise to create jobs and bolster the economy. What are the measurable benefits? How will river traffic be affected by these shipments? How will road traffic be affected? What is the danger zone? This would be a typical use case.
In the latter, more general case, it first helps provide a common foundation for debate in civic society about pressing issues, secondly it provides an educational tool for people to even learn about an issue, thirdly it provides a way for people to develop an intuition about the balancing points of the complex systems that we live within, and fourthly it levels the playing field between different stakeholders. Overall it helps provide a new kind of decision making in civic society as an alternative to and qualitatively different than making donations to a campaign, running for office, voting occasionally, protesting in a parade or writing letters to the editor.
If your idea were to become a reality, who would benefit the most and how?
(you mean aside from lil ol me? i’m an ecosystem too you know - i mean 80% of me isn’t even me but a kind of fauna ecology that is colonizing the scaffolding of my body… think about that! next time you kiss your girlfriend/boyfriend/dog )
There are literally thousands and thousands of small grass roots humanitarian and environmental organizations with the will, the time and the power to effect change, all they need is better information so that they can work in concert with other organizations.
There are ordinary citizens with day jobs who can only pay partial attention to issues of economy, environment or policy.
There are corporate interests that develop resources and that have to work within what we permit them to do.
There are civic planners and legislators, hired to make good decisions for us, and to raise some decisions for vote.
These subscribers would also be the primary beneficiaries. A predictive modeling tool that brought to bear the resolving power of hundreds of human minds would make many formerly intractable problems quite shallow.
At the end of the day even in most kind hearted decisions we make, we just don’t have the understanding we need to end up at the right conclusion. Issues in civic planning, land use, human and natural impacts are simply too complex right now. In some cases the side-effects of new bylaws even dominate over the intended consequences. What we need is a collaborative predictive modeling tool with buy-in from all participants that gives us all a better gauge of the right direction to go in.
What are the initial steps required to get this idea off the ground?
(money money money support hacking hacking hacking more money more hacking more support more more more )
I see a staged set of deliverables. First simulate only a single watershed in coarse detail. Take SRTM elevation data for Oregon and or California and model water flow including the water evaporation/rain-fall cycle, turbidity, temperature, gravity, aquifers. This is implemented as a cellular automata and each of the ‘agents’ in the simulation will be having an effect on the fabric of the space defined by the automata. After this simulate salmon, a key indicator species that brings protein to inland forests and that bridges two very different ecosystems - land and ocean. With water and salmon we have the beginning of a cross domain model.
At this point we want to make the service public. A problem with many academic environmental models is that they are used to support a thesis or paper but are not persistent and not public. We want to provide a durable, persistent, open website where any user can come in and instance a new simulation or clone an existing simulation, rewind, fast forward, set granularity of computation, upload new agents, include or exclude agents defined by other users.
After this point the goals would no longer be technical but promotional. Initial value would be best for specialized communities trying to model very specific problems, and as they add new agents the tool will have increasing utility and value for wider audiences and more complex interactions.
Describe the optimal outcome should your idea be selected and successfully implemented. How would you measure it?
(this was a trickier question - because I’m not into many of the metrics people create - but at least it was a chance to voice thoughts about better metrics )
The optimal outcome would be that people are happier, healthier and fit into their working landscapes better. Local watersheds and ecosystems are healthier. A generation of SimCity kids emerge who have an intuition about the space of civic planning and who respect and work with their world better.
Narrow empirical measurements can be done by interview; asking people using the tool if the tool is more or less effective at helping them negotiate a given civic issue between different stakeholders.
More broadly although the tool is intended to improve quality of life - I think it’s probably dangerous to select for human welfare as a baseline measurement - I see the Peter Singer calculus of “harm minimization” to be degenerate. Rather some measure of creativity, freedom and joy should be used as the measuring stick. A good metric would also be to measure local before and after regional biodiversity on the theory that biodiversity probably means more people have more choices and are more happy.
Usability studies and interviews should be conducted to get feedback on the effectiveness of the tool. Quality of life should be measured and defined by a multiplicity of factors including general human health, free time, quality of local physical environs. We also want to measure civic participation in terms of awareness of local issues and any feeling of satisfaction in having a voice in local issues.