Testability.com
Reference
The Information Source for Systems Testability and Diagnostics

Basic Dependency Modeling Terminology

Event – A phenomenon that follows and is caused by some previous phenomenon. One of the elements in a dependency model.

Agent – An entity—physical or conceptual—that is capable of producing a certain effect. One of the elements in a dependency model.

Dependency – The relationship between a given element (event or agent) and another element upon which it is causally dependent. The term is also used to refer to an individual element in its role as a cause of a given element (e.g., "agent x is a dependency of event z").

First-Order Dependency – A dependency that represents a direct causal relationship (e.g. if agent x produces event z, we say that "x is a first-order dependency of z").

Nth-Order Dependency – A dependency that represents an indirect causal relationship (e.g. if agent w generates event x which triggers agent y to produce event z, we say that "w is an nth-order dependency of z").

Dependency Statement – A representation of the first-order and/or nth-order dependencies of a given event. Individual dependency statements almost never exist in isolation, but rather appear as constituent statements within a dependency model.

Full-Order Dependency Statement – A dependency statement that represents all dependencies that are directly or indirectly responsible for a given event.

Dependency Model – A representation of the causal relationships between events and the other elements (events and agents) that enable those events. Dependency models are comprised of multiple dependency statements that collectively represent the causal interrelationships within a system, device or process.

Dependency Modeling – The process of generating a dependency model by separately defining its constituent dependency statements. In many applications, the individual events and agents within a dependency model may be represented at such a low level that the model is most easily manipulated using higher-level constructs and abstractions (such as those used in Passive-Active Flow Modeling and Test Overlay Modeling).

Logic Modeling (LogMod)  – The name used by Ralph A. Depaul, Jr. to refer to the causal modeling technique that he invented in the 1950s and which would later become known as dependency modeling.

Initial Event – An event within a dependency model that is represented by a dependency statement that contains no dependencies. The initial events within a dependency model correspond to the inputs of the system, device or process that is represented by that model.

Terminal Event – An event within a dependency model that is not listed as a dependency within another dependency statement. The terminal events within a dependency model correspond to the outputs of the system, device or process that is represented by that model.

Upstream – A term used to describe the topological relationship between a given element (event or agent) in a dependency model and another element that is an nth-order dependency of the given element (e.g., "an initial event is upstream from those events and agents that it affects").

Downstream – A term used to describe the topological relationship between a given element (event or agent) in a dependency model and another element that contains the given element among its nth-order dependencies (e.g., "a terminal event is downstream from those events and agents that have an affect upon it.")

Single-dimensional Dependency Model – A dependency model that contains a single set of dependency statements and which is used for modeling a system, device or process in which conditional causality does not come into play. Not to be confused with single-signal modeling.

Multi-dimensional Dependency Model – A dependency model that contains multiple, alternative sets of dependency statements (or a single set of dependency statements that are modified dynamically) in order to support conditional causality. In a diagnostic application, a multi-dimensional dependency model can be used to represent test assymetry and changes to dependencies that result from system state dynamics. Not to be confused with multi-signal modeling.