An Empirical Study on the "Usage of Not" in Real-World JSON Schema Documents (Long Version)

TL;DR

This study investigates the practical usage of negation in JSON Schema documents, revealing that negation is indeed used in real-world schemas and follows consistent patterns, which could inform future schema design practices.

Contribution

It provides the first large-scale empirical analysis of negation in JSON Schema, demonstrating its practical relevance and identifying common usage patterns.

Findings

Negation is actively used in real-world JSON Schemas.

Usage patterns of negation are stable and consistent.

Negation has potential to evolve into standard design patterns.

Abstract

In this paper, we study the usage of negation in JSON Schema data modeling. Negation is a logical operator that is rarely present in type systems and schema description languages, since it complicates decision problems. As a consequence, many software tools, but also formal frameworks for working with JSON Schema, do not fully support negation. As of today, the question whether covering negation is practically relevant, or a mainly theoretical exercise (albeit challenging), is open. This motivates us to study whether negation is really used in practice, for which aims, and whether it could be - in principle - replaced by simpler operators. We have collected the most diverse corpus of JSON Schema documents analyzed so far, based on a crawl of 90k open source schemas hosted on GitHub. We perform a systematic analysis, quantify usage patterns of negation, and also qualitatively analyze schemas. We show that negation is indeed used, following a stable set of patterns, with the potential to mature into design patterns.