Glossary

ASCII - (American Standard Code for Information Interchange). The character set on PCs, VAXes, and UNIX machines.

Column - a (vertical) array of data values that represent the different values of one attribute over many instances (or rows) of that attribute. Used in a relational context to describe the concept of data element. Loosely equivalent to a 'field' on the mainframe.

Data Element - the smallest unit of information that can be understood (or perceived) by an end user. (Durrell definition). A logical data element becomes a column in a relational context and a field in a record-oriented context.

EBCDIC - (Extended Binary Coded Decimal Interchange Code). The character set on IBM and plug-compatible mainframes.

Entity - The existence of a thing as contrasted with its attributes. In the context of entity-relationship modeling, an entity is a business object that is eventually represented as a table.

Entity attribute - A characteristic of an entity that is to be recorded for all instances of an entity. In the entity-relationship modeling process, entity attributes become data elements and then table columns.

Entity instance - a specific occurence of a defined business object that is eventually represented as a row in a table.

Field - the smallest logical unit of data in record-oriented (e.g., mainframe application) environment. Corresponds to a column in a relational environment. Both are physical instances of a data element.

File - a collection of records in a record-oriented environment. Corresponds to one or more tables in a relational environment.

Hierarchical Database - a database that is organized in a tree structure, in which each record has one owner. Navigation to individual records takes place through predetermined access paths. (GG definition.)

Integration - the process of unifying entities across subject areas and time. Views which perform the integration are built as part of this process.

Key - the columns (in a relational environment) or fields (in a record-oriented environment) that together uniquely identify a row or record.

Network Database - a database (like IDMS) that is organized in multiple tree structures, in which each record has one or more owners. Navigation to individual records takes place through predetermined access paths.

Normalization - the process of reducing a data structure to its simplest form. Basically, normalization is the process of functionally decomposing, or breaking down into modules, a complex data structure until it is in its simplest form. A normalized data structure minimizes data redundancy and data element coupling, and maximizes data element cohesion. Normalization is an attempt to achieve the most flexible and stable data structure possible, so future maintenance of the data structure can be reduced. Normalization is the process of developing a data structure that is in the nth normal form (often referred to as "highest" or "best" normal form). Normalization is to data as structured design and programming is (sic) to process. (Durrell definition).

Record - in a file-oriented environment a record is a collection of fields that normally contain information about one individual or entity. A record may contain many segments so that it corresponds to one or more rows in a relational environment.

Referential integrity - checking that values in one table that are intended to be used as keys to another table are indeed found in the second table.

Relational database - a database that is made up of normalized tables. Many logical links between tables are declared, but links can be made that were not predetermined when the database was originally designed.

Row - a (horizontal) array of data values that represent the different values of each attribute for an individual instance of a data entity. This term in a relational environment corresponds to a record in a file-oriented environment.

Segment - a physical group of data elements in a mainframe record that is repeated many times and which describes a subordinate entity of some sort. A segment usually becomes the row of a distinct table in a relational context.

Table - a two-dimensional array of data composed of rows and columns. Some tables are actual physical entities managed by the Data Base Management System (DBMS) while others are views that join physical tables when the view is referenced.