1.5. Building the Data Dictionary

The models provide convenient places to record the system’s information as it becomes available. For example, at the beginning of the Project in Chapter 1.2 Start with the Context, you learned that a programme supplier tells Piccadilly about new programmes. You added the data flow NEW PROGRAMME to the context diagram to show this communication between your system and the outside world. Because the new programme information has to be remembered, you also added the entities PROGRAMME and PROGRAMME SUPPLIER to the data model.

As soon as you learn the contents of the NEW PROGRAMME data flow, you define it in the data dictionary. Whenever you define a new data flow, you often discover new entities and relationships, so it’s back to your data model to record them. As you understand the attributes of the entities, you complete the cycle by recording them, too, in the data dictionary.

Every time you discover something new about Piccadilly, you continue to add to the appropriate models. Each time you add something, the model reveals some new facet of the system, or indicates other areas that are fertile ground to investigate. Eventually, the whole of the system will be discovered and recorded.

Index cards make a convenient repository for your data dictionary entries. Make one entry per card and file them alphabetically in a box. Alphabetically tabbed cards make it easier to find your entries. Or, if you prefer, write your entries on regular paper—it’s just a bit harder to retrieve your entries later.

Using a word processor or database to store your entries makes your dictionary easier to manage. However, please don’t spend too much time setting up elaborate schemas to store your dictionary. Remember that index cards are perfectly adequate for the number of entries you’ll need in this book. If you are using a CASE tool to record your models, use the built-in dictionary.

We discourage you from using automated data dictionaries for this case study for the simple reason that you don’t have to write enough entries to justify using automation (we will provide most of the dictionary for you). However, when you move on to analyzing your own projects at work, you’ll have to write hundreds, perhaps thousands, of dictionary entries. The tasks of building and maintaining the data dictionary for large projects both need and benefit from automated help.

Read the following description of part of the activities of the Programme Transmission Department. Think about the data this department uses, because you are going to write some data definitions.

Figure 1.5.1: The data flow NEW PROGRAMME must contain enough information for Perry Vale and his staff to decide whether to buy the new programme being offered.

Let’s look more closely at these data. The programme always has a name, which gives some idea of its content. Programmes are categorized by type. For example, it may be a first-run film, sporting event, documentary, talk show, old movie, or one of the many other types of programmes. When the supplier describes the programme’s contents, he sometimes provides a list of all the performers appearing in the programme. Some programmes include the names of the producer and director.

Another factor in deciding the suitability of the programme is its duration. For instance, Abel Gance’s epic film Napoleon runs for five hours. There is no point in buying it unless you have a five-and-a-half-hour transmission slot available (Perry has to allow for the commercial breaks).

The supplier always tells Piccadilly the programme price. As Piccadilly is a British company, the price is always stated in pounds sterling.

Figure 1.5.2: These are the various rates for 30-second spots. There are similar pages for 10-, 20-, 40-, 50-, and 60-second spots.

A brief explanation is needed here to get you started. First, note that the day of transmission is important, and the rates vary depending on whether the spot is transmitted on a weekday, a Saturday, or a Sunday. The days are broken into segments that match the times when people watch television.

The headings—“Fixed,” “Broad,” “ROD,” and “ROW”—specify the degree of moveability of the spot. A fixed spot will be transmitted in a designated break, on a designated date. (Note that “date” and “day” have different meanings in this context.) Advertisers buy fixed spots if they wish to take advantage of the audience of particular programmes. For example, to have an advertisement shown in a break during Coronation Street (Britain’s highest-rated soap), you would have to buy a fixed spot.

A broad spot can be moved within its segment, but must be transmitted on the specified date. For example, a broad spot sold into the prime-time segment on a weekday can be transmitted in any break between 18.00 and 22.40 (6 p.m. and 10:40 p.m. in the U.S.). Broad spots are moved into breaks with lower ratings if another advertiser is willing to buy a higher-priced fixed spot.

Run-of-day, or ROD, means that the spot can be transmitted in any segment with the same price, on the designated day. For example, an ROD spot sold at £1800, could be shown between 16.30 and 18.00, or between 22.40 and 23.40, on the designated day. Run-of-week, or ROW, allows the spot to be moved to any same-priced segment within a designated week. If you bought an ROW spot for £1200, it will be shown sometime in the afternoon segment on any one of seven days.

Piccadilly naturally prefers to sell at the highest price, whereas the advertisers want to stretch their budgets by buying the cheapest spots they can. So the advertisers buy at the lowest price they think will neither be preempted nor moved to the poorest allowable break.