Triples-Lang is a software language designed for AgenticAI Agents to create their own tools, integrated with a RDF database seamlessly and communicate with other AIs with ease.
As its name implies, it is focused on interacting with a RDF database and the syntax follows a subject, verb, object (SVO) paradigm as much as possible. The SVO paradigm was chosen because it links directly to a Resource Description Framework (RDF) database meaning the returns for queries can be directly executed. Also, as I mentioned in an early post, LLMs are well suited parsing natural text to SVO and a non-LLM parser can be easily developed for deployment environments where LLMs are not an option. Triples is an explicit language reducing potential confusion and enabling it to be read out loud for interfacing verbally with a robot. However, its primary motivation is for an AI – not a human – to naturally connect to a RDF database.
Semantic Triples (a.k.a RDF Triples) are a common data structure used for knowledge base creation.Triples codify a statement into a sequence of three entities.
We do not directly use RDF but a more flexible format, similar to conceptual graph interchange form (CGIF), worked better for most AgentAI. which call Portable Semantic Commands (PSC).