Abstraction

Abstraction is the process of retaining essential features while setting aside irrelevant details. It makes complex things easier to describe, reason about, or use.

Information hiding: A design principle that conceals internal details behind a stable interface. It separates what an abstraction exposes from how its implementation works.
Relevance: The quality of being connected to a particular question, purpose, or context. A feature is not essential in the abstract; its relevance depends on the task.
Abstract data type: A data type defined by its operations and behavior rather than its implementation. It specifies what a data structure does while omitting how it is built.
Concretization: The process of making an abstract idea or representation more specific and detailed. It moves in the opposite direction, restoring particulars that abstraction leaves out.
Cognitive load: The mental effort required to hold and process information during a task. Abstractions reduce the amount of detail that must be considered at once.
Idealization: The deliberate simplification of a subject by assuming features that are not literally true. It shows how abstraction can preserve a useful pattern by omitting complications.
Feature (machine learning): A measurable property or variable used as input to a machine-learning model. Features are selected representations of data, not the full underlying objects.
Application programming interface: A defined set of rules through which software components communicate. An API exposes capabilities while hiding internal implementation details.
Reductionism: The approach of explaining complex phenomena through simpler components or lower-level processes. Reductionism seeks explanatory parts, while abstraction selects features suited to a chosen purpose.
Black box: A system understood through its inputs and outputs while its internal workings remain inaccessible or ignored. A useful abstraction can become a black box when hidden mechanisms matter.

Linked from 13 pages