The text read off it
[page 68]
Level numbers are used to describe the way in which a body of data is organized. Basically, level numbers are assigned to items of data to show their relationship to other items of data—or, in other words, to show the structure of a record. Any number from 1 to 99 can be used. All data description entries must be assigned level numbers.
In general, each item is considered to be a subdivision of the last item preceding it which has a lower number. Figure 2 shows how a typical series of files, records, and fields might be organized, using the familiar method called outlining. The file structure is shown by the use of indentation, each item being considered a part of the last item above it which is indented to a lesser degree.
The technique of indentation, in other words, is a visual way of showing level. It may be used in the "Data Name" columns, but it will have no effect on the processor itself. However, since it helps to identify the various levels visually, indentation may be useful in clarifying the file structure when the program is listed.
For comparison, two additional columns have been provided at the right in Figure 2. These show the data classification of each item, together with hypothetical level numbers such as might be assigned to a file structure of this kind. It should be pointed out that entries for files and groups of files are not actually used in the data description.
It is obvious from the outline that each item from EMPLOYEE NUMBER through LOCATION is a part of PAY RECORD, and that each item from FICA through HOSPITALIZATION is a part of DEDUCTIONS, YEAR TO DATE. Had the principle of indentation not been used, the reader might still determine these relationships by examining the level numbers in the right hand column, following the rule that each item is part of the next item above with a lower number. The processor, of course, will ignore any indentation and will store the data in accordance with the level numbers.
It is not necessary that level numbers be assigned in consecutive order, although it is done that way in Figure 2. The items at level 02, for example, might have been assigned level 04, or any other convenient number, as long as it was greater than 01. Similarly, the items at level 03 could have been given any other number as long as it was greater than the number of the next higher classification. In fact, it is often useful to skip numbers when they are initially assigned, to allow for possible regroupings or insertions at a later time.
The reader will also note that each item at the record level and below represents a kind of data, not a specific item of information. Thus, although there will be only one file called EASTERN REGION SALES FORCE, within that file there will be many individual units called PAY RECORD, and each of these will contain information of the same general character and format, as specified by the names of the fields within it. The purpose of the data description is to give information about each of these kinds of data. The data description should be thought of as a "pattern" which the files will follow.
It may be helpful to consider a second method of showing how data may be organized. Figure 3 shows how the same payroll file might be represented in the form of an organization chart.
This chart demonstrates one important fact: Each item of data is, or may be, related to other items above and below it. In other words, the data is organized "vertically"—no item is related directly to any other item at the same level, except through an item at a higher level. One result of this is that if a particular item is of the same kind as other items in the file, it may be identified by reference to the item above it in the organization structure. Thus, the name PAY RECORD does not single out any