Alain Colmerauer and Philippe Roussel created the first Prolog system in 1972 at the Artificial Intelligence Group of the Faculty of Sciences of Luminy. Their work emerged from Aix-Marseille II University in France during a period when logic programming was still an emerging field.
The name Prolog came from Philippe Roussel, who chose it based on his wife's suggestion as an abbreviation for Programming in Logic. This early version relied on Robert Kowalski's procedural interpretation of Horn clauses to reconcile declarative knowledge representation with procedural methods popular in North America.
SLD resolution serves as the primary method used by Prolog engines to determine query success through resolution refutation. When multiple clause heads match a given call, the system creates choice points allowing alternative execution paths that continue until all possibilities are exhausted or success found.
ISO/IEC 13211-1 published in nineteen ninety-five standardized existing practices among numerous Prolog implementations clarifying previously ambiguous aspects. Three corrigenda followed: Cor.1:2007, Cor.2:2012, and Cor.3:2017 maintaining the standard through ISO/IEC JTC1/SC22/WG17 working group oversight.
IBM Watson utilized Prolog within its DeepQA software framework alongside Java and Apache UIMA for natural language processing tasks. GeneXus low-code development platform incorporates Prolog technology focused around artificial intelligence applications while TerminusDB open source graph database implements core functionality entirely in Prolog designed for collaborative knowledge graph building and curation.