IS 240 -- Systems Analysis and Design

Assignment for Week 4: Growth of the technology.

Traditionally, systems analysis and design has presupposed a narrow design space of mainframes and terminals. Designing a computer system was a routine matter: map the information flows, model the data, and you're almost done. Not so any more. The design space is much larger now, and yet we cannot free ourselves of the stale habits of the past until we get a concrete sense of what that new design space is like. This week we will gather raw material. To take some examples, Moore's Law says, in effect, that computer processors grow 100 times more powerful every ten years. Memory densities grow at a comparable rate, and mass storage densities grow even more quickly. Computer networks are growing rapidly more capacious, even if it's not obvious that the average household will enjoy these rapidly expanding bandwidths anytime soon. Wireless data services have been slow to take off, but they are grabbing hold ferociously in Europe, and we can now hope for general-purpose wireless packet networking in the foreseeable future if not right away.

This is just a rough sketch; your assignment is to paint a picture of the design space in ten years that is more complete and more compelling. This is library work, and Web work, and it will take you into the trade press. It may require you to talk to technical people to check your understanding of things. It wouldn't even hurt to read the ranting of futurists like Ray Kurzweil. The technology forecasting organization Institute for the Future (IFTF) is respectable; their best-known spokesman is Paul Saffo. And it will make you think about how to present such information usefully. A 100 GHz processor sure sounds cool in the abstract, but what does it mean for us? How can we make the numbers imaginable and useful? Now, I understand that this assignment is, strictly speaking, impossible. Nobody knows what the world will be like in ten years. So what you are really reporting is guesses and predictions, and the experts' are not much better than yours. You are limited, obviously, by the strengths and weaknesses of the public literature. If you have access to specialist reports in this area, by all means please share. But you will not necessarily find predictions framed precisely in terms of "ten years", or 2010, or whatever. Work with whatever predictions you can find, and guess.

It is helpful to distinguish between two kinds of predictions. One kind is quantitative. The fact that we can extrapolate from past quantitative improvements in processor speeds and the like is most unusual: prediction is rarely so reliable or mechanical. Find out the historical growth curves, and then calculate accordingly. You will find many discussions in the literature -- even in publications such as the New York Times -- of these growth curves, and speculations about how long they will last, and whether new inventions will enable us to run ahead of them.

The other kind of prediction is the qualitative organization of architectures and standards. These are remarkably easy to predict as well, but for the opposite reason: because they change so slowly. A standard such as Jini (which allows networked devices to advertise their capabilities to one another), Bluetooth (a spontaneous wireless networking standard meant for things like home appliances), or TCP/IP (the protocols of the Internet) can languish in laboratories for a decade or more before the necessary critical mass can be established to for them to "take off" and be widely adopted in the real world. We can therefore safely guess that most of the standards that will be widely adopted in 2010 already exist today, at least in nascent forms in the laboratory, and that few entirely new concepts will enter the pipeline in time to change the world by then.

You have two jobs: present some reasonably solid information, and build a bridge between information and imagination. Design starts with imagination, and good designs will only get built if they can first be imagined. What is Bluetooth, really? What can you do with it? What could you do in a world that was saturated with Bluetooth-compatible devices and services? Okay, so databases will be 1,000,000 times bigger (or whatever the best estimate is). What does that mean? What will fit in that kind of database? Ask around for references that might address the issue. Actually, you have three jobs: the third job is to provide the class with a bibliography, Web links, handouts, and other resources for those who want the details.

Remember that you are not supposed to kill yourself doing this assignment; we only want you to do as much as you can do in the time available. Why are we asking you to hunt down this material rather than us doing it ourselves? Well, we've already done a lot of it ourselves, though surely not as much as you will. Our real reason is that we want you to confront the literature in this area for yourself. A designer in the real world must constantly predict and track various developments that might soon intersect with his or her own projects. If you commit to using a certain standard in your product, for example, you acquire a strong interest in whether that standard will survive at all, how quickly it will be adopted, what directions it will take, and so on. You may be motivated to attend standards meetings, express opinions on heavily-trafficked Web sites, and contribute code to open source projects. The point, then, is that designers are embedded in a social world as much as a technical world. By predicting the future you are joining into the group project of building it. The forms of imagination that you develop will be part of a collective imagination, not least because you have an interest in helping other people to imagine the future in the same way that you do.

Your presentation can take many forms. This material is usually presented in a sober way with PowerPoint slides with statistics on them. If that's your way of bridging information to imagination, go for it. Otherwise, rethink. What picture is emerging from your research? Define the organizing principle of that picture and let it organize your presentation.

Remember to document your work activities while doing this project, and to reflect on the process of research and design. Your experiences will become important raw material for later in the course. Observe your collaboration in all its aspects. Experiment. Watch yourself use different media, various information resources, and so on. What are the real practical problems? How does your group process fit (or fail to fit) with the rest of your life? What sorts of tools would make the whole thing easier? Your goal is to make life better for all of the interdisciplinary teams in the world.

First group: An, Hoffman, Lauruhn, Miller, Sorrell.

Raw computing power. The quantitative predictions here are easy enough because everyone seems to believe in Moore's Law. But so what? What could we do with 100 GHz microprocessors anyway? If steering wheels cost 1% as much as they do now, it would make no difference to the cost and functionality of cars. Perhaps people would burn steering wheels for fuel, but otherwise the world would be much the same. You will encounter a lot of loose talk about how computers will become as intelligent as human beings, or 1% as intelligent, or whatever. This is misleading at best, given that computers work completely differently from people. So find out, what applications of computers, currently in the pipeline, would actually benefit from a 100-fold increase in computing power? Good places to look include speech recognition, simulation, and financial derivatives. But even then, just saying "speech recognition" is only of limited help. Again, so what? What difference does it make to the information services we can deliver?

Second group: Childers, Lippincott, Napper, Tsay, Vasquez.

Databases are obviously crucial for designing information services. How big will they be in ten years? What kinds of information will be in databases that aren't commonly in databases now? Will databases of video be useful in ten years, or is that more like twenty years? How much of a database will someone be able to carry? How many databases will there be in the world, and will we be able to interact with large numbers of them at the same time? Take the conventional scenarios about the whole Library of Congress being available in one big database: are those scenarios true, or is it more complicated than that? What kinds of metadata will be widely standardized, adopted, useful, and consequential in ten years? What will that let us do that we can't do now? One could obviously write a long book about these questions, but we want the concise, riveting, worldview-changing seven-minute version.

Third group: Cleary, Gazan, Hessel, Hunt-Coffey, Plutchok.

In Europe, the wireless revolution is now. Tell us about WAP, the Wireless Applications Protocol. What is it, really, and what are some of the services that are already happening on it? A few sentences will convey the general idea, but it's important to say them because the United States is behind in this area. What can we do with WAP in ten years? Or are there other standards for delivering wireless services that will grow to maturity in that time? Will wireless digital networking really be universal by then, at least in the advanced economies? What kind of bandwidth will we routinely get over wireless connections by 2010? Will we have video cell phones? (Never mind for now whether we would want them.) What can we do with wireless network connections to cars? That sort of thing. In this area you'll find that the ratio of cheap predictions to reliable analysis is quite high. See if you can separate the serious stuff from the hype.

Fourth group: Dodge, Khoshoo, Hektor, C.-N. Wang, P. Wang.

What will the Internet be like in 2010? Will the whole telephone system be running over the Internet? Will millions of people be sending video signals at the same time? Will the Internet fail to scale up and collapse into a chaotic mess? How many devices will be attached to it? Think particularly about the tension between the quantitative aspects of the network (bandwidth, for example), which change quickly, and the qualitative aspects (applications standards, for example), which do not change quickly. The Internet has been growing exponentially in number of hosts, but the Web works pretty much the same way that it worked in 1995, or indeed in 1990, and e-mail works pretty much the way it worked in 1990 as well, unless you count the spread of MIME, though it works much better than it did in 1980. Keep in mind that our goal is not to dive into the technical details, gnarly as they are. Rather, we want to characterize the design space. What does a systems designer need to know to imagine what useful information services will be practical in 2010? That's the focus.

Fifth group: The mystery group, aka everyone else.

Investigate technologies that are related to physical space, especially GIS (geographic information systems) and GPS (global positioning satellite). Many potential information services benefit from knowing where people are, or from tracking vehicles or goods. Many other services deliver information that is indexed by physical space: weather, environmental data, travel-related information, services related to security and safety, and so on. Will we be able to build a GPS tracking device into a cell phone? How wide a variety of GIS resources will be available? All of this stuff exists now. The question is whether it is cheap, practical, widely used, and generally ready for prime time.