COMTRAN Reconstruction

F28-8043 · June 1960

Page 78

PDF page 84 of 124. Download the scanned PDF

The page as scanned

The scan of PDF page 84 of F28-8043
Select the scan to open it at full size.

The text read off it

[page 78]

To call for any one item of data, the programmer must write as many subscripts as are necessary to identify the particular level. Thus, STATE (3) would call for the third item called STATE, DISTRICT (3,2) would call for the second DISTRICT field within the third STATE item, and CITY (3,2,6) would obtain the sixth CITY field in the second DISTRICT of the third STATE. Subscripts are always written in descending order of level.

Mode (Col. 36)

The term "mode" refers to the method by which data is represented, such as the binary mode or the binary coded decimal mode. For the purposes of the Commercial Translator, the mode used in the arithmetic units of the system is considered to be the system's "internal" mode, although the system may be able to read data in other modes.

Column 36 is used to show whether the data being described is prepared in the internal mode for the system being used. In that case, the letter I is punched in this column. Data punched in an "external" mode is represented by the letter E. Specific information on the modes available for each system will be provided in the publications covering the processors for the various machine systems.

Since a body of data may contain information in more than one mode, the mode is specified at the lowest level of data organization—i.e., at the level where the "field pictorial" is shown. (See the "Description" columns.) Thus, if a record at the 01 level contains two fields at the 02 level, one of them in the internal mode and one in the external, there would be no point in specifying a mode for the record itself. The mode (or modes) of a larger unit is therefore a consequence of the mode specified for each of its components.

Justify (Col. 37)

The term "justification" is used in printing and writing to mean the alignment of a margin. Thus, the text of this manual is both "left justified" and "right justified." In data processing, the term has a similar implication, since it is usually necessary to specify that data be justified if the programmer wishes it to be printed out with an aligned margin.

Actually, however, justification means a good deal more than this. Specifically, it refers to the placement of the data in a unit of storage. In many systems, data is stored in machine "words" of fixed length. Very often, therefore, a particular item of data will be shorter than the space allowed for it, and it may be necessary, if alignment is to be preserved, to provide a means of filling the unoccupied spaces with non-significant characters, such as zeros or blanks, depending on the system. On the other hand, it may be desirable to fill all available space with data, so that more than one item—or parts of more than one item—may be stored in a given machine word. This is known as "packing."

It is extremely important to recognize that justification specified for an input item of data—i.e., an item to be entered into the system at object time—describes the item as it already exists. It does not change the justification; it is used in informing the system where to look for the incoming data. Justification specified for other items, however, actually controls the placement of the data. It instructs the system how to handle the data when it operates on it internally or prepares it for output.

The effect of specifying justification is as follows: If left justification is specified, the data item is stored, or will be stored, so that its left-hand character is placed in the left-hand position of the next available machine word. This may mean that the right-hand portion of the preceding word is left unoccupied. If right justification is stipulated, the data item is stored, or will be stored, to the right in the next available machine word, leaving unoccupied whatever portion of that word is not required at