Welcome back. Today we will be talking about
the issue of scale as it relates to the topics of this course. Some of the ideas I will be
presenting in this lecture come from a book that I wrote in 2010 called “Greening through
IT” that discusses some environmental issues and information technology. This came out
through MIT Press a few years ago. The core argument in the book is that environmental
issues tend to take place on broad scales of time, space and complexity. They occur
across thousands of miles and hundreds of years and are vastly complex interacting among
many different domains, and the time horizons of these issues are different from the ones
that humans are potentially more familiar with. So, there are things like sea level
rise that take place over 50 or 100 years rather than an earthquake which is pretty
instantaneous. There is a distinction between urgent versus being important, but a lot of
things humans are good at responding to are urgency, not necessarily importance. This
is a distinction that I think is critical. So, sea level rise is one example. So here
we see a graph from 1880 to the year 2000. This only demonstrates a few centimeters worth
of sea level change, and yet the fact that it is this inexorable rise and that it is
forecast to continue up leads to some pretty significant challenges where large portions
of human inhabitation will be underwater at a certain point. This appears to be something
that is an ongoing trend. While these environmental issues are broad in time, space and complexity,
humans evolved with regard to a relatively narrow set of problems. We evolved to deal
with narrow things in time, space and complexity. The kind of feedback loop that evolution allows
for is only on the scale of human bodies. So, in terms of time at most tens of years,
but often shorter times then that. In terms of space usually within arm’s reach, but
certainly no more than a couple of years walking distance. In terms of complexity our brains
aren’t necessarily good at synthesizing vastly different domains all at the same time,
so relatively narrow in terms of complexity. One demonstration of this, there exists a
system called Mechanical Turk. This is an Amazon product where you can go online and
you can pose questions to people. There is something like 10,000 people who are there
willing to answer your question for a dime or 50 cents. I went and did a study, just
sort of a back of the envelope calculation with this where what I asked was what is the
relevant time scale for a variety of different topics? I asked them to say what was the time
scale for danger, for hunger, for information technology, for financial issues, and for
environmental topics. The outcome when I asked people this was that danger tended to happened
on a scale of about 1.2 hours, hunger was about 4.2 hours, information technology was
3.2 months, financial things were 7.2 months, and environmental topics were 8.1 years. So,
here we see the difference between the scale of various different topics. On this graph
it is very hard to see because 8.1 years is so high compared to 1.2 hours, and I wanted
to show you another way to look at these same data. This is looking at the same graph on
a log scale, a logarithmic scale, and here you can see the relationship between them
a lot more clearly. A logarithmic scale is based on powers of 10. So, the first line
is the 1s, and the next line will be times 10, and the next line will be times 100, the
next line will be times 1000, and so this allows you to see exponential variables, exponential
growth, rather than linear variables better. One thing I would encourage you to do in terms
of thinking about scale is to go look at this XKCD comment which I think does an interesting
job of pointing out the relative scales of a lot of different things in the scope of
human experience. So, looking at this distinction between large scale environmental problems
and relatively small scale human problems, there is an important role for technology
that across human history technology has bridged the gap between large scales and small scales.
So, the core argument in my book is that environmental issues take place over big time scales and
space scales and human horizons are relatively small, but that throughout human history technology
has allowed us to make connections between these small scales and larger scales, as well
as putting small scales and even smaller scales, and that information technology in particular
is well suited for this effort. I also wanted to make a point about books. Ultimately they
are written by people. I’m just a guy and I wrote this book, and parts of them are wrong.
Sometimes they are even wrong by the time they are in print. So, there are various aspects
of the “Greening through IT” book that when I read them I’m like oh, I wish I hadn’t
written it quite like that. Hopefully they provide some sort of a useful perspective,
but when you read a book note that it is just written by a person who had some opinions
and they thought to write them down. Often they are an expert in the field, but that
doesn’t mean they are necessarily perfect in their understanding of a topic, or that
their understanding doesn’t change over time. There is another potential way to look
at the idea of scale and this is looking at first order effects, second order effects,
and third order effects. So, first order effects are relatively narrow in their perspective.
So, efficiency for example as a first order effect seems like a really good thing. It
gets the same work done with less energy. That seems like it would be clearly sustainable,
but looking at a second order effect, these are broader indirect effects. So, if you do
something more efficiently than you save some money, and what you do with that money, you
may go out and spend it on a trip to Hawaii or somewhere and that then means that the
money you saved you might have other environmental issues. These are second order effects. Third
order effects are broader still. These are broad-scale civilizational effects. Things
like the existence of a culture of consumption where people just get in the habit of using
a lot of stuff because it is so freely available. So, an interesting finding in the area of
indirect effects comes from a thing called the Jevons Paradox. So, in 1865 William Stanley
Jevons was doing a study of steam engines in England. What he was expecting to find
in his study was that as steam engines became more efficient the overall use of coal would
go down, but what he found was actually the opposite. As steam engines became more efficient
more and more coal was being used. That’s not because any given steam engine for a given
period of time was using more coal, but it was that more steam engines were being used.
Steam engines were being used in more industries, and they were being used for more tasks and
for longer periods of time. So, even though they were becoming more efficient, the overall
use of coal went up during that period of time. That’s why it is a paradox. More efficiency
led to more coal use instead of less coal use. So, President Obama’s executive order
from a few years ago calls for increased energy efficiency as part of the Federal Government’s
“integrated strategy towards sustainability” but there is a question. Does efficiency necessarily
lead to greater sustainability? Efficiency certainly leads to greater economic prosperity,
at least very likely, that it allows you to do more stuff with the same resource and so
it leads to prosperity, but does it lead to sustainability? That depends a lot on some
of the indirect effects. So, it makes perfect sense why a government would be exactly in
favor of efficiency because of economic prosperity, but this complex relationship between efficiency
and sustainability is also an important part of this topic. Another aspect that impinges
upon this is the time lag on indirect effects. So, whereas direct effects take place immediately,
indirect effects often take weeks or months or years or decades to ripple through all
of the interconnected systems and have their effects. So, the time discounted value of
all direct and indirect effects might be another way to think about it. What is the value over
the course of time that these indirect affects have? So, various changes like increased efficiency
might buy us time to address some of these other issues as well, but it also may just
delay the point of which we feel like we as a civilization need to think about some of
these sustainability issues. In the field of science scale also plays an important role.
Different sciences look at different phenomena at very different scales, and this one sometimes
will migrate within that same discipline across a variety of different scales. So as an example,
traditional ecology research has been a relatively small scale phenomenon. It was a field research
activity where individual researchers would go out into the wild and engage in harsh conditions
with notebooks and pencils slogging around in the mud. As an example of this there is
a professor who recently retired a couple years ago, retired from UCI, named Lynn Carpenter
who studied Costa Rican Restoration College. They are trying to turn fields that had been
over-farmed back into usable farm land and ultimately potentially into rain forest. Her
field site was full of muddy boots and milk crates full of plants, and that was the way
in which she engaged with her research and she was doing some really important and interesting
work in that capacity. But, over the last few years, well actually the last few decades,
ecology has become much more of a big science where there are many different researchers
who are trying to integrate across the different studies and doing cross domain research. Rather
than think about one species in one location, they might be thinking about many different
species in the same location, or the same species in many different locations, the complexities
among all of these different projects rather than one particular species in one location.
Ecoinformatics is the field that looks at how information technology can be brought
to bear on ecology as a science. So, using information technology to enhance ecological
research and change the way that ecological research can happen. But, here we have a challenge.
How do you bridge the gaps among thousands of different scientists with millions or billions
of different data points all collected in different ways? One person is collecting data
using one set of techniques, and somebody else is using a different set of techniques.
How do you bridge this gap? How do you merge across different species, and the geographies,
and different habitats, and data quality? So for example, citizen scientists are regular
citizens of a given country who are being enlisted to do things like count the number
of birds that they see out of their back window each day. The data quality you get from that
may not be as reliable as a PhD student or some other researcher who is being paid to
sit there and count the birds, and yet it allows for a much broader set of activities
when you can list thousands of citizens rather than just a few researchers. The reason I
wanted to address all of these topics of scale is to cause you to think a little bit more
about the role of scale in your projects and the way you think about the topics of this
class, because environmental issues and sustainability issues have a lot of scale aspects involved
in them and hopefully IT can be brought to bear productively in addressing some of these
scale issues and over the course of this quarter we would like you to think very much about
how scale affects all of the different topics of the class. Thanks very much


Leave a Reply

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