2.3. A Variety of Viewpoints

Viewpoints are a justified distortion of the reality being represented. Using viewpoints, analysts can build models that include only as much information as they need to see. This doesn’t mean that models are a false representation of the system. It does mean the models are more usable because they show what the analyst needs at the time, and such models don’t burden analysts with details that can be delayed or shown in another model.

A well-known example of justified distortion is the map of the London Underground shown in Figure 2.3.1. Anyone who has visited London has probably seen, if not used, the map of its Underground. Each of the “tube” (subway or train) lines is in a different color running horizontally, vertically, or at a 45-degree diagonal. This map uses a viewpoint. It shows the way the stations are connected, and discounts their actual geographical location. Why? Consider the people using the map. They want to travel from one tube station to another. What matters to them is the way their two stations are connected. By straightening out the lines, the map designer makes it easy for travelers to see which lines and which interchange stations must be navigated before they arrive at the intended destination.

Figure 2.3.1: This representation of the London Underground* is displayed throughout the trains and stations, and frequently appears on postcards and T-shirts. The map distinguishes the lines of the network with different colors.

*Copyright © 1933 by London Regional Transport. Reprinted by permission.

In reality, the Underground is not at all like the colored map. A topographical map showing the true layout of the lines would be almost unrecognizable to Londoners. The topographical map represents reality, but it has not been used since the other map was produced in 1933. Why not? The reason is simply that it shows a viewpoint that is of little use to Underground passengers. When you are a hundred feet below the streets, you have little interest in knowing whether you are passing under the Old Bailey or Nelson’s Column.

Figure 2.3.2: Topographical map of the London Underground. This map shows the correct geographical location of the lines and stations.

The stylized map in Figure 2.3.1 takes great liberties with the geography of London, but the distortion is entirely justified. By filtering out unnecessary information, it presents a viewpoint that is far more useful for its intended audience.

Today’s information systems are big, and more and more systems serve large areas of the users’ business. In the past, it was sufficient to build a computer system that produced a single report. Those days are gone. Today’s computers are more powerful and capable of accommodating elaborate systems. Increasingly sophisticated users are expecting integrated systems that bring together the different activities of their businesses, as well as higher levels of specialization and functionality within each application.

Figure 2.3.3: Yesterday’s inventory report is no longer enough. Today, the computer system must cover inventory control, warehousing, sales, purchasing, and more.

For example, if you live in a large city, chances are that each day, you are exposed to several thousand pieces of advertising material. The average American sees more than eighty television commercials each day. Sunday editions of the New York Times have carried up to 350 pages of advertising. Italian and French drivers negotiate a forest of eye-catching billboards erected beside their roads and on conspicuous buildings. Add to this the magazines, newspapers, subway posters, shop displays, handbills, and packaging that everybody sees. Each day, we are exposed to an amazing barrage of advertising material urging us to buy something or other.

Figure 2.3.4: Did you take careful notice of all the advertising you saw or heard yesterday? Were you impressed by all the claims of bigger, better, newer, or brighter? You probably ignored most of them or gave them at best a cursory glance before continuing on your way.

What is it that keeps you from absorbing this oversupply of advertising material? It is your viewpoint. You filter out the advertisements you don’t need or don’t want to see at the moment, and look at those you do. You ignore the advertisements for coffee if you are a tea drinker. You skip the ads for bottled water if you like tap water. Most men disregard the pitches for women’s cosmetics and most women ignore the electric shaver commercials, except at Christmas and birthdays. Your viewpoint is saying, “I don’t need to see that now.” You keep your sanity by filtering out unneeded information.

Similarly, systems analysts must filter information to be effective. As systems get larger and extend to more areas of the users’ business, the number of people who have knowledge of part of the system increases. More people means a wider variety of attitudes, skills, and vested interests confronting analysts. To deal with the increasing number of user groups, analysts can respond by building system models that show the viewpoints of greatest interest, or most relevance, to each audience.

Analysts also need to filter out what is not needed at the moment. When analyzing a large system, they need to distort the reality by modeling the viewpoint of the system that contributes most to their understanding. The viewpoints we’ve found most successful for our own projects, and the ones we’ll introduce in this book, are the current physical, essential, data, and new physical. Let’s meet them.

Current Physical Viewpoint

Figure 2.3.5: The current physical model focuses on a system’s current implementation. While it shows the system requirements, the model also includes the processors that do the work. This view usually shows departments, people’s jobs, existing computer systems, machines, and so on.

The current physical viewpoint is used in a limited way at the beginning of a project to establish the system’s context of study, and to provide users with recognizable models. Users are initially more comfortable if they see models that show the people and machines currently performing the tasks. It also helps the analyst to identify problem areas, to get to know the sources of the system’s information, and to assess the impact of the future system. We discuss this topic fully in Chapter 2.8 Current Physical Viewpoint.

Figure 2.3.6: The essential viewpoint, which shows only the business policy of the system and ignores the machinery that does the work.

Figure 2.3.7: A data model of the system. This abstract viewpoint looks solely at what the system has to remember.

Unfortunately, life is not so simple as to allow analysts to build a complete model of one viewpoint, verify it, and then move on to the next one. The information required for each viewpoint, as well as the users who supply the information for the viewpoint, will appear at random times.

As the project analyst, you cannot expect to get all the information you need when you need it. You have to be prepared to capture it whenever it becomes available. As a result, you will typically have several models, each showing a different viewpoint, in various stages of completion. This does not hinder the analysis effort. Your models, though incomplete, enable you to continue gathering the requirements.

For most of the time, you’ll model the Piccadilly system’s essence. Toward the end, when the essential viewpoint is largely complete, you will select the new operating environment and construct a new physical viewpoint of your implementation recommendations.

The Trail Guide will lead you to the viewpoint chapters when you need them and to the chapters that discuss the various models that represent the viewpoints.