¤ Identify fields to be encoded:
Fields to be encoded are determined either by the existence of an algorithm that derives values from meaning (e.g., abbreviation, encryption), or by consideration of all attributes classified as designed (i.e., an attribute invented to represent some other attribute or relationship). The most common designed attribute is the numeric identifier of an entity that has other means of identification, such as a name.
¤ Define the code structure:
Select or identify the following:
* Method of encoding (e.g., mnemonic, encryption, binary choice, serial number, arbitrary)
* Method of decoding (e.g., relationship/linkage, table look-up, associated field)
* Zones (e.g., sub-codes within the full code)
* Data type (e.g., numeric, alphabetic, alphanumeric)
* Responsibility for code values (e.g., external organization, department, data administration)
* Whether associations between code values are supported
¤ Assign code values:
For codes with a finite set of meanings, assign a code value for each meaning and define associations with other codes and meanings, if allowed.
A set of values may be derived from the values of the domain of the attribute represented (e.g., abbreviation of months of the year).
A special case of encoding is the representation of a basic set of objects in a multilingual environment. Here, the starting point may be a set of meanings for one of the languages.
¤ Identify code control procedures:
Consider the style of code, and the responsibility for code values. Serial numbers may be allocated by system or clerical procedure. The procedure may be a current system procedure.
Procedures may need to support the life cycle of entities identified by the code (i.e., creation, identification, authorization, transfer of responsibility or other relationship or discontinuance).
» Creates deliverable D4.7.5 Preliminary Data Code Structure.
Previous : A4.3.1
Activity Overview : A4.3
Stage Overview : Design Stage
Overview : Table of Contents