The text read off it
[page 71]
one unique kind of item, but SALES FORCE PAY RECORD distinguishes one kind of item from any item designated PRODUCTION FORCE PAY RECORD. Similarly, since in this illustration there are two SALES FORCE files, each can be identified by reference to the next higher level. Thus, a particular kind of item might be known as EASTERN REGION SALES FORCE PAY RECORD.
The use of a higher name to identify a lower one is known as "name qualification," and a name so qualified is known as a "compound name." This is an important method of identifying individual items which do not have names unique in the program. It is true that in this illustration the qualifying names are those of files. However, the principle is applicable internally, within files, whenever it is necessary to distinguish between fields having the same name. (See the discussion of compound names beginning on page 15.)
Level numbers are not actually attached to the data in the sense that an employee number is part of a pay record. They are used to instruct the processor to perform certain technical functions which need not concern the programmer. Essentially, they are used before the actual data is read into the system, as a means of preparing the system to receive it. Once the data description has been written, the programmer need no longer concern himself with level numbers unless, owing to changes in the data or the program, a new data description should become necessary.
Type (Col. 25-30)
Columns 25 through 30 are used, when necessary, to show that the data being described is of a certain special type. If these columns are left blank, it will be assumed that the remainder of the particular entry describes a data field or group of fields. The type codes which may be used in these columns are the following:
RECORD
COND
FUNCT
PARAM
REDEF
COPY
LABEL
Each of these is discussed in the following pages.
RECORD
This type code shows that the data being described is a record and is therefore accessible by GET and FILE instructions. This is equivalent to identifying an item of data as an input/output record. Each record named in the data description must also be named in the environment description, as specified in the publication covering the processor for each particular system.
COND
The type code COND is used to show that the data referred to is one of the possible conditions which a conditional variable may assume. In the discussion of conditional expressions in Chapter 2, it was pointed out that a conditional variable is the name of a field which will contain, at different times, any of a number of different values, depending on conditions existing in the data. Each of the values that may be placed in the field is a "condition."
In an example used in Chapter 2, the name MARITAL.STATUS was given as the name of a conditional variable. This name refers to a specific field reserved in storage into which values representing conditions will be entered. Typical conditions for this