systems

About

Syndication

Categories

Archives

Moving Outside the Device

design, systems

March 19, 2004, 03:31 PM

I've been the proud owner of an Apple iPod for a while now. Although I'm happy with my purchase, I do have some issues with the little device. These issues can be summed up in a sentence: it doesn't really talk to any of my other devices. Not that its alone. My devices in general are a rather antisocial bunch.

I generally listen to my iPod on my walk to and from work. When I get home, I remove it from my pocket and plug it into the computer to recharge the batteries. Assuming I haven't pushed "pause", why not have the song I was listening to immediately start up in iTunes? I generally want to at least finish listening to the song that was currently playing when I arrived home. But my iPod only talks to my computer to synchronize my music collection. Which is wonderful functionality, but since they're talking to each other already, why not go a little bit further?

Likewise, when I get into my car I often want to listen to my music collection; I have a cd player but choosing and transporting all those little plastic discs is a pain. Since my whole collection is on my iPod, why not provide a car radio that I can slip the device into (Kenneth calls this the "iPod ker-chunk slot") so that bringing my music with me is dead simple?

This is what distinguishes experience design from product design; experience designers consider the interfaces and interactions that appear throughout the whole system the customer is a part of. It's not enough to create products that provide point solutions but don't integrate with the rest of the customer's world; the smart consumer wants his problems solved by his purchases, and these problems span his interaction with any particular device.

The good news is that I think Apple, at least, gets this philosophy. It remains to be seen where they'll go with it, and who will follow.

[ Permanent Location ]

A Systemic Problem with Software

design, software development, systems

December 24, 2003, 11:33 AM

One of the core themes you'll find running through this weblog is an approach to analyzing problems from a systemic perspective, that is, an approach that examines the structure of the larger systems these problems exist to gain a deeper understanding of their causes. Recently I've been thinking about systemic problems in software design, and I wanted to articulate the one that's especially jumped out at me.

First off, lets assume a world in which all our dreams have come true. Human-centered design is the norm rather than the exception. Efficient production and distribution models have arisen for software design and development, and the market is perfectly in tune with the needs of its end-users. Sadly, the following problem would still plague us.

Here's the crux of it:

1. Designing good software requires maintaining a strong focus on a small, well-defined user group. Cooper makes this argument, and it's pretty well-known to anyone who has done serious interface design work. Different people have different needs, and any interface that tries to be too general, that tries to do everything for everybody, will only wind up being mediocre (at best) for many, and completely unusable for the rest. Feature creep and the complex, inelegant interaction designs that inherently accompany it will see to that. This is partially why Word and Windows suffer from so many usability problems; they are trying to satisfy too many disparate user groups.
2. But people need to work together, so their software must work together too. Users need to exchange documents, send each other messages, work collaboratively on the same artifacts, etc. If they software they use doesn't support this, it cannot fulfill their goals.
3. If people have different software, this creates an interoperability problem. Different software systems, especially those made by different companies, are notorious for causing problems with interoperability, from the old issues with opening Wordperfect documents in Word to modern problems with viewing web pages in Safari that were created for Internet Explorer.
4. There will always be tension between focused design for usability and generalized design for interoperability. Often, the result of this situation is that natural monopolies get created in the software market, such as Microsoft. One company produces a "one size fits all" solution that everyone uses to avoid the problems with interoperability, but which fails to fully satisfy anyone's needs.

So what can be done about these problems? One option is to explicitly design software to interoperate with other software, which I've discussed before. But it's hard to design two pieces of software so that they communicate together effectively, even when you have full control over both of them. And designing for connectivity with all the software your customers might potentially be using quickly creates an exponential complexity problem.

The obvious solution is to define software standards and design to conform to them. However, we know that usability is not screen-deep, so these standards may put limits on the usability of the final products. Ideally, the standards should be developed while following a user-centered design process as well, but here we come back to our original problem; standards by definition must take the needs of many user groups into account, which is inherently difficult to design for. What usually happens is that standards arise from the experience of many product development efforts, and attempts at standardization that occur too early fail as vendors ignore them in favor of adding proprietary capabilities to their software to fulfill their customer's needs.

I don't see any obvious way out of this messy process, dictated by the structure of software design and development. My guess is that firms must do the best they can in the particular situations they find themselves in, and work to enhance the standardization process so that less time is wasted discarding inappropriate standards.

[ Permanent Location ]

Empirical Research on the Reasons for Free Riding

society & sociology, systems

September 04, 2003, 08:51 PM

For CSCW this week, we had to read an empirical research article relating to free riding / social loafing and summarize it for the class. I chose an article that studied the reasons why people tend to free ride as a long-term trend in Prisoner's Dilemma-style situations, since I'm pretty interested in game theory economics (although not ordinarily so interested that I'm willing to slog through academic journal articles on the topic). I'm posting it here because I found the results interesting enough that I believe a higher-level summary of the experiment is worth disseminating. Just in case you're interested in the real thing, the reference for the article, in APA-style, is: Andreoni, J. (1988). Why free ride? Strategies and learning in public goods experiments. Journal of Public Economics 37: 291-304.

In "Why Free Ride? Strategies and Learning in Public Goods Experiments", James Andreoni is concerned with determining the cause of "free riding" in game economics experiments. Free riding is the term used to describe behavior where an individual acts in his own self-interest at the expense of the group because this will result in greater benefits to him, even though the optimal behavior is for all members of the group to act in the group's interest. For example, the classic experiment in free riding studies involves forming groups of five subjects and giving each member of the group 50 "tokens". Tokens can only be redeemed for cash by investing them in one of two funds: an "Individual Exchange" fund returns 1 cent to the investor and nothing to his group members, whereas a "Group Exchange" fund returns a half cent to the investor and everyone else in the group. Note that the game involves the same general situation as the Tragedy of the Commons; the rational individual will choose to invest in his Individual Exchange fund since that always returns 1 cent to him (and he'll also gain from any Group Exchange investments made by his teammates), but the best strategy for the group as a whole is for all members to invest all their money in the Group Exchange, since that returns 2.5 cents to each group member. It's also important to note that none of the participants are allowed to communicate about the game, so they cannot, for example, all agree to follow the optimal strategy.

Andreoni notes that all experiments of this type tend to result in a convergence on free riding behavior when the game is played repeatedly, whereas the behavior is evenly divided between free riding and public interest when the games are "one shot" deals. He describes two theories for why this occurs: the "learning hypothesis", which postulates that individuals don't initially understand the game, but, over time, they learn that free riding is the optimal strategy for their personal gain. The "strategies hypothesis" goes further and claims that some individuals learn the optimal strategy but seek to further maximize their games by occasionally contributing to the Group Exchange fund to prevent others from learning the game or realizing that they are playing rationally so that their groupmates will contribute more to the Group Exchange fund. However, at some point near the end they will "bail out" and stop contributing to the group fund, which explains the end-game free riding behavior.

To test these theories, Andreoni ran an experiment where a control group played the normal repeated game (he calls this group the "Partners"), whereas a variable group played a version of the game where the participants in each group were randomly shuffled at each iteration, so that effectively each player was just playing a series of "one shot" games (he calls this group the "Strangers"). This eliminated the effects of the strategies hypothesis (it makes no sense to try to "psych out" your groupmates if you are assigned new groupmates after each iteration) so that only learning could explain any observed behavior in the variable group. Moreover, both groups were subjected to an unexpected "restart" where the game continued for three additional rounds after they were told it would end. This was to isolate the learning hypothesis; if learning was primarily responsible neither group should be affected since both had already learned the game. If either group was affected, then something else had to be going on.

Andreoni summarizes his results as a series of six observations:

1. Investments in the Group Exchange was greater by the Strangers than the Partners in all rounds. This is evidence against the strategies hypothesis, which claims Partners would always be greater since all players invest more in Group Exchange when they are fooling or being fooled.
2. The percent of Partners who free ride is greater than Strangers in all the rounds, with the greatest difference in the last round. This is more evidence against the strategies hypothesis.
3. The Partners give the least in the last round, but most have still not reached the free riding equilibrium (the amount they would contribute as the number of rounds approaches infinity). Once again, more evidence against "strategies".
4. The Strangers give more than the Partners in the last round. If learning is all that's going on, then the Partners must be learning much faster than the Strangers, which is hard to accept. This suggests that learning is not solely responsible for the trend towards free riding.
5. The Strangers appeared to be only temporarily affected by the restart.
6. The Partners are affected, and return to high levels of investment in the Group Exchange in the restart. This appears to have a lasting effect. Remember that the learning hypothesis predicts that neither group will be affected if it is the sole cause of the free riding trend.

Andreoni summarizes by asserting that his experimental data contracts both the learning and the strategies hypotheses as accounting for the tendency towards free riding.

Finally, he concludes with a discussion on what theories and new experiments may account for free riding behavior. For example, he points out that it is possible that participants may have learned the optimal behavior for the "one shot" game, but don't yet understand the dynamics of the repeated game. This would help explain the fluctuations after the restart, since participants may not recognize that the continuation of the game should not affect their behavior. He also suggests the examination of theories that look at "non-standard behavior" that consider the reasons subjects make their decisions rather than just the characteristics of the equilibrium. For example, subjects may get non-monetary pleasures from cooperating. Or they may be enforcing social norms on cooperation by trying to punish those who don't contribute to the Group Exchange. Or participants who do worse than the expected may play more cautiously and not contribute as much to the Group Exchange. In conclusion, Andreoni calls for an empirical examination of a broader range of alternatives such as these.

[ Permanent Location ]

The Tragedy of the Commons

politics, society & sociology, systems

September 03, 2003, 12:38 PM

Last night, I read a paper for CSCW on "The Tragedy of the Commons", which I believe is the seminal paper that applied this concept to modern political science. If ever there was an argument that knocked down the "Invisible Hand will fix everything" theory of capitalism (still popular among some of the more overzealous libertarians), this is it.

The paper is very good and pretty readable, so I'd strongly suggest you read it. But since I know most of you won't, here's a summary of the argument, as applied to the problem of overpopulation:

1. As the number of people increases, the amount of available resources on this planet per capita will decrease.
2. Since population, unchecked, grows exponentially, we will soon reach a point where there is only enough resources per capita for bare subsistence, so no one will have sufficient resources to enjoy life (in reality I'd argue this is an unlikely scenario; it's much more likely that a minority will have sufficient resources and the majority will not have enough to survive, but the important point is that either scenario is bad news).
3. Therefore, overpopulation is a big problem.
4. However, from the perspective of the rational self-interested individual actor (whom the invisible hand is supposedly guiding), the cost of having another child is nearly entirely beneficial since the burden of overpopulation is distributed evenly to all humans and thus is negligable to him, whereas the cost of not having another child is entirely detrimental to him.
5. Therefore, the rational self-interested individual actor will conclude that he should always have another child.
6. And, of course, so will all the other rational, self-interested actors out there.
7. Therefore, the overpopulation problem will continue to get exponentially worse until the scenario described in (2) occurs.
8. Sucks to be us.

This is an insight that exposes a core problem confronted by modern economists and other social thinkers. It proves that a blind, simplistic trust in markets is misguided, and that the world is much more complicated than most laissez faire propoganda would have you believe.

This isn't to say that markets are bad, of course, just that they need to be fixed sometimes. The interesting debates are over how to do this; how to balance freedom with responsibility. But in order to engage in this debate, you first have to accept the reality that it must occur.

Hardin claims that attempts to convince people through education that they must voluntarily reject the course of action that benefits them the most for the sake of society at a whole are inherently flawed and unworkable. He argues:

If we ask a man who is exploiting a commons to desist "in the name of conscience," what are we saying to him? What does he hear? — not only at the moment but also in the wee small hours of the night when, half asleep, he remembers not merely the words we used but also the nonverbal communication cues we gave him unawares? Sooner or later, consciously or subconsciously, he senses that he has received two communications, and that they are contradictory: 1. (intended communication) "If you don't do as we ask, we will openly condemn you for not acting like a responsible citizen"; 2. (the unintended communication) "If you do behave as we ask, we will secretly condemn you for a simpleton who can be shamed into standing aside while the rest of us exploit the commons."

Similar to the Prisoner's Dilemma, the system ensures that the people trapped in it cannot escape through cooperation unless they can be convinced to all trust one another, an unlikely scenario.

The only alternative, Hardin argues, is "Mutual Coercion Mutually Agreed Upon". Essentially this involves stamping out the commons and replacing it with systemic structures that either carve out portions and allocate them to individuals ("private property") or artifically make the costs of abusing the commons much higher than the benefits (fines, incarceration, etc). Essentially this entails giving up freedoms, but its a mutual sacrafice that humans agree to on their own volition. Of course, Hardin skirts the issue of what exactly those people who choose not to enter into the agreement are supposed to do... I don't like this conclusion, but so far I must say his argument is convincing.

The core idea, however, is that we as intelligent actors have to alter the system as a whole to work for our long-term best interests rather than against them. But this introduces problems of its own. Systems are complex, and it is hard to predict all the real effects changes will have. I've mentioned the "Law of Unintended Consequences" before. Hardin himself brings up the phrase "Quis custodies ipsos custodes?" or "Who shall watch the watchers themselves?". Whenever you introduce changes, there have to be mechanisms in place to enforce those changes. Usually this requires enforcers, who are human too and subject to the same rules as all the other humans in the system. You can't expect enforcers to be angels; they will look out for their own self-interest just the same as everyone else will.

Is it even possible to design perfect systems that can really address all these issues? Or is this just a cleverly disguised technical solution to a "no technical solutions problem", which Hardin deplores in the first few paragraphs? T.S. Eliot once said "It is impossible to design a system so perfect that no one has to be good". Where, then, lies the hope for lasting betterment?

[ Permanent Location ]

Ayole's Water and Technology As a Catalyst for Change

charity, society & sociology, systems

July 11, 2003, 10:50 PM

First off, a quick apology for the recent slowdown in posting frequency. I've taken on way too many projects this summer and thus haven't had as much time to update this journal as I'd like.

Today I partially taught a lesson in the TCinC class I'm co-teaching with Matt. Matt has put together the lesson plans for most of the sessions we're redoing for Joe's class, and today he had us show a video to the students called The Water of Ayole.

The Water of Ayole is a video used by the Peace Corps to communicate the important concept that technology must be introduced into organizations along with the skills and systems necessary to maintain and develop that technology. The video itself is about a program to introduce water pumps into many third-world african villages. The basic problem this program is trying to solve is that many villages are reliant on water sources that are unsanitary, disease-ridden, far from the village (and thus take a lot of time and effort to collect), etc. This lack of potable water is responsible for the deaths of thousands of Africans every hour. To solve this problem, the program funded the construction of water pumps that are capable of drawing water from deep underground that is pure, potable, and close to the village (and thus much easier to gather).

Unfortunately, although the intentions of these programs are good, the execution was flawed. Here's what generally happened:

1. The government came in and funded the construction of the water pump. The villagers were overjoyed.
2. A couple of years pass. The pump breaks.
3. The villagers complain to the government, but the government doesn't have the money to maintain the pumps for free.
4. The villagers attempt to raise money to fix the pumps, but they are poorly organized. The money is never collected and used appropriately.
5. The women of the village go back to the time-consuming task of gathering the unsanitary water they drank before. Disease and death returns.

The exception to this rule is the village of Ayole. When the interventionists brought the pump to Ayole, they didn't simply build the pump and leave. Instead, they helped teach a mechanic who lived in the village how to repair the pump. The villagers themselves formed committees who were responsible for discovering problems with the pump and raising money for repairs. They even engaged in communal farming to help raise money for pump maintenance as a common good for the village.

Clearly, this video demonstrates the importance of organizational change when incorporating new technology. But I took another message away from it as well. I've always believed that true, lasting change can only occur if it happens in the hearts and minds of the people, and thus technology cannot, by itself, create change. However, technology can be a catalyst for change. The villagers in Ayole remarked that their community had grown stronger as a result of the pump, that the presence of the common good brought them together, even in ways that were not related to pump maintenance. They now took a greater interest in each others' affairs, and were more willing to help fellow villagers in need. Of course, this change did not come from the pump per se; it was the result of the organizational change that swept through the village as a response to the need to maintain the pump. But the pump undoubtedly spurred this change. Introducing a new technology can, when other conditions are right, tip the scales in favor of a new, better social order. And I think that's pretty exciting, especially for a budding technologist who hopes to do some good in this world.

[ Permanent Location ]

On Cooperation and Understanding Others

society & sociology, systems

May 29, 2003, 03:50 PM

An important skill to cultivate if you wish, like me, to make a difference in the world is the ability to put yourself in the position of the people you must work with so that you can understand clearly what their motivations are, what external forces are influencing them, and why they act the way they do and make the choices they make. Here's why.

Psychologists have known for a long time that although we humans have a tendency to blame our own failures on external forces beyond our control ("I bit Bill's head off because I had to sit through rush hour traffic this morning, but I'm not an irritable person", "I did a bad job on the paper because I didn't get enough sleep last night, but I'm still an excellent researcher"), we have a tendency to blame other people's failures on their personal character ("Jack's a jerk; he yelled at me for a friendly joke!", "This is the third crappy paper Jill's put out; she's a lousy researcher!"). But this tendency blinds us to the fact that all people work within systems; they are all part of a larger whole that includes the other people they work with, the physical environments they live in, the technologies they use, etc. These systems define what these people must do (if they wish to maintain social and material standing), what they want to do (if they wish to improve their lot in the system), and what they cannot do (if they wish to avoid censure or some other punishment).

When someone opposes your goals you might be tempted to think that person is uncaring, or incompetent, or evil, when in reality they simply have external influences affecting them that you may be unable to see. More often that you might think, people with good intentions are hamstrung by systems that are not designed to support those intentions, or even designed to actively oppose them. It takes careful thought and empathy to fully understand such a situation; a surface analysis never turns up everything.

If you can't understand the larger systems that exist behind the people you encounter and how these systems are influencing them, you will be quick to demonize the people around you, to write them off as foolish because that's simpler that trying to understand the complex reality. But when you must work with these people, this only serves to foster a competitive environment where everyone is out for themselves. The only way around your coworkers is through them.

On the surface, this may seem faster. Surely fighting for what you know is right is faster than talking and compromising and working together (at least if you win). But the assumption here is that your ideas will work "out of the box", that the world would be a much better place if only everyone listened to you. This is patently false, no matter who you are. No one's ideas will work right off the bat because no one is in a position to see the whole world for what it is. There will always be complications, always a need to iterate and improve. Cooperation seems slower only because you are actually seeing these iterations in progress; you're constantly making small mistakes and small corrections as everyone checks and balances each other. But if you chose to compete, then each iteration requires a whole process of taking sides, building armies, fighting, killing (hopefully metaphorically), etc., just to have a bunch of new ideas put in place that frequently turn out to need just as much iterative improvement as the old ones did. If you think long term, if you keep your eyes focused on the ultimate goal, then antagonism is almost always an unnecessary distraction that gets in the way of real progress.

All this isn't to say that people's values are not important, but values alone don't tell the whole story. I can value intellectual openness, but if my bread and butter is coming from selling my work by the copy then I'm not going to embrace eliminating IP laws. I can value saving the environment, but if I feed my family with money I earn from a logging company, I'm not going to support stricter deforestation regulations unless you can convince me my needs will be taken care of. If you refuse to see the system people must work within, then you won't win them over to your side. You don't understand them. And no cooperation can occur without understanding.

In the end, all real progress occurs through cooperation. If you can't understand the needs of the "other side", then everyone's goals will suffer, regardless of the outcome.

[ Permanent Location ]

A Cost-Effective Problem Reporting System

design, systems

May 22, 2003, 01:55 PM

Earlier today I was in one of the CMU bathrooms near where I work and I noticed that one of the urinals was continuously flushing. I thought of contacting facilities to report the problem, but I was busy, I have no idea what the number for facilities is, and I was suffering from the "Somebody Else's Problem" effect.

Later on, I was thinking that a problem reporting system to fix this breakdown would be pretty easy to implement, especially since CMU has a campus-wide wireless network (which seems to be spreading; one day in the not-to-distant future wireless may be near-ubiquitous). All it would require is a small embedded device with a single "big red button" interface that could be placed in areas where equipment frequently breaks down, people need assistance, or some other situtation where one person needs to notify a geographically distant person that a problem has occurred in the area. When someone pushes the device's button, it uses an embedded wireless card to send a notification over the network to a central server that processes the notifications and routes them to the people in charge of maintenance for the area so they can quickly respond. See the diagram below.

I can think of several extensions to the system to customize it to particular situations, such as adding the ability for the problem reporter to leave a quick voice message describing the problem as they hold down the button to give more context, or a fingerprint-recognition system on the button to identify the problem reporter so the support person can contact them for more information or discourage pranksters from pushing the button as a joke. But in its simplest form, this system solves a practical problem, is easy for support people to maintain, and is cheap to implement if the wireless infrastructure already exists (I can't imagine the problem button devices would be difficult to mass produce; they seem pretty simple to me).

Someone should further develop this idea and market it. I have to start making friends with some business students.

[ Permanent Location ]

Pattern Sighting in Management Habitat

design, patterns, systems

April 05, 2003, 02:12 AM

I was working with Matt a few days ago on our TCinC independent study, and he showed me part of a book on Systems Diagrams, which are a component of systems thinking as described in The Fifth Discipline by Peter Senge. What struck me as particularly interesting about these things was how similiar Senge's concept of system "archetypes" is to patterns.

For those of you haven't heard of them, patterns are generalized solutions to a set of problems that occur in a particular discipline. They were first described by Christopher Alexander, a building architect, in his 1977 book A Pattern Language. Alexander's architectural pattern language included patterns such as "Courtyard" and "Good Materials". The Gang of Four (Gamma, Helm, Johnson, Vlissides) applied this notion to the design of software systems in their 1994 book, Design Patterns, which is where I first heard of them. They also happen to be one of the ten thousand things I'm very interested in.

In software systems, one common problem is that you have a component that performs a function you want to use, but it doesn't conform to the interface you need it to. Here's an example from a fictitious drawing editor application:

In this situation, we have an inheritance hierarchy whose root is Shape that has a "getBoundingBox()" method. We want to subclass Shape with a class that implements a line of text, and we already have a component, TextView that we'd like to reuse. However, TextView doesn't conform to Shape's interface. So we create TextShape, which implements Shape's interface just by calling the corresponding method(s) on TextView. This solution is then generalized to the Adapter pattern:

These drawings were adapted from diagrams in Design Patterns.

The Fifth Discipline is not related to software design, but instead is a business / management / project planning text that attempts to analyze work environments to locate the "feedback loops" or parts of the work process that must be examined if sustainable organizational learning is to occur. Senge provides archetypes of problem systems that commonly appear in organizations so you can try to fit your system to one or more of the archetypes to better understand what sort of solutions to apply.

Note how similar this is to the concept of a software design pattern; in both cases the assumption is that an abstract problem in the domain has an abstract solution structure that can be distilled by looking at many successful solutions found in practice. Then this solution structure can be adapted to fit local conditions to solve specific problems.

I wonder how other disciplines apply patterns or pattern-related concepts to solve common problems. In a sense, patterns are nothing new; we're just using induction to build abstractions, then deduction to apply those abstractions. It's how the scientific method works, and how humans have learned and thought for millennia. Yet making this way of thinking and encoding knowledge explicit seems to have a power of its own. I wonder where it can take us.

[ Permanent Location ]

© 2003-2008 Rob Adams ()

Last Modified: Friday January 25, 2008

Unless otherwise indicated, all original work appearing on this web page is licensed under the Attribution / No Derivitive Works Creative Commons license. Please before using any page content in a manner not covered by this license.