This paper introduces a terminological framework for describing transaction-oriented recovery schemes in database systems. It defines key concepts such as materialized database, propagation strategy, and checkpoint, enabling the classification of arbitrary implementations from a unified perspective. The paper also presents a classification scheme for logging techniques, which are precisely defined using these terms. It shows that these criteria are related to important recovery questions such as speed, scope, and the amount of redundant information required. The primary purpose of the paper is to establish an adequate and precise terminology for database recovery, which is a complex area with many conceptual and implementation-related issues. The paper discusses the concept of a transaction, which is central to database recovery. A transaction is a sequence of database operations that must be treated as a single unit. It has four key properties: atomicity, consistency, isolation, and durability (ACID). These properties ensure that transactions are isolated from other transactions and that their results are durable. The paper also discusses different types of failures that must be anticipated in database systems, including transaction failures, system failures, and media failures. It describes the recovery actions required for each type of failure, including transaction UNDO, global UNDO, partial REDO, and global REDO. The paper also introduces the concept of a mapping hierarchy in database systems, which describes the different levels of abstraction used to represent data. It discusses the storage hierarchy, which includes the physical storage of data and the different levels of abstraction used to represent it. The paper also discusses the different views of a database, including the current database, the materialized database, and the physical database. These views are important for understanding how recovery actions are applied to a database. The paper concludes by discussing the importance of recovery in database systems and the need for a precise terminology to describe the various recovery techniques. It emphasizes the importance of understanding the different types of failures that can occur in a database system and the recovery actions required to handle them. The paper also highlights the importance of the ACID properties in ensuring the reliability and consistency of database transactions.This paper introduces a terminological framework for describing transaction-oriented recovery schemes in database systems. It defines key concepts such as materialized database, propagation strategy, and checkpoint, enabling the classification of arbitrary implementations from a unified perspective. The paper also presents a classification scheme for logging techniques, which are precisely defined using these terms. It shows that these criteria are related to important recovery questions such as speed, scope, and the amount of redundant information required. The primary purpose of the paper is to establish an adequate and precise terminology for database recovery, which is a complex area with many conceptual and implementation-related issues. The paper discusses the concept of a transaction, which is central to database recovery. A transaction is a sequence of database operations that must be treated as a single unit. It has four key properties: atomicity, consistency, isolation, and durability (ACID). These properties ensure that transactions are isolated from other transactions and that their results are durable. The paper also discusses different types of failures that must be anticipated in database systems, including transaction failures, system failures, and media failures. It describes the recovery actions required for each type of failure, including transaction UNDO, global UNDO, partial REDO, and global REDO. The paper also introduces the concept of a mapping hierarchy in database systems, which describes the different levels of abstraction used to represent data. It discusses the storage hierarchy, which includes the physical storage of data and the different levels of abstraction used to represent it. The paper also discusses the different views of a database, including the current database, the materialized database, and the physical database. These views are important for understanding how recovery actions are applied to a database. The paper concludes by discussing the importance of recovery in database systems and the need for a precise terminology to describe the various recovery techniques. It emphasizes the importance of understanding the different types of failures that can occur in a database system and the recovery actions required to handle them. The paper also highlights the importance of the ACID properties in ensuring the reliability and consistency of database transactions.