Frame-QL: An Introduction

Abstract

Frame-QL is a query language for declaring a governed analytical result. The requester names theanalytical objects to be returned, declares the anchor at which the result frame should exist, and statesthe conditions that shape it. The requester does not prescribe the relational operations or physicalcommands used to produce the result.This design follows from the ontology of the Theory of Data. A measure is a stable governed family,such as revenue, inventory level, or order count. A member of a measure is one governed realization ofthat measure at a particular analytical location under a member contract. A frame is an assembly of co-anchored fields at one output anchor; in the fully governed case, every field is a member of a measure. AManifold records the measures, root measure members, family laws, populations, boundaries, evidence,lineage, and physical bindings required to adjudicate the class of measure-member derivations it supports.Because that knowledge is governed outside the query, a requester can ask for the customer-month metricof the revenue measure without writing the path that constructs it. The planner resolves and certifiesthe request; a trusted engine may then realize it through SQL, dataframe operations, a cache, an API,or another backend.Frame-QL has two anchor mechanisms because a query has two structural levels. The query-level ATclause is mandatory and declares the one final output anchor of the frame. Inside metric expressions, @ascribes an analytical anchor to a subexpression. That ascription can govern the input metric consumedby an operation and, in a nested expression, the intermediate output metric that becomes the nextoperation’s input. This is not just syntax: for anchor-sensitive operations, the input location can be partof what the requested quantity means.This paper develops the central idea of Frame-QL: the output frame is the query. It explainswhy rows, tables, joins, and GROUP BY are not primitives of the request interface; how measure identitydiffers from a physical column; how ambiguity becomes clarification rather than silent choice; and whyFrame-QL gives AI agents a useful interpretive role without granting them authority over analyticallaw or database execution. The Frame-QL Manual, Second Edition remains authoritative for syntaxand shipped behavior. The Theory of Data, Version 5.0, is authoritative for the broader ontology andcertification model.