Reversible matrices, labelled graphs, and labelled 2 structures

[LLyaudet]

Hello,

I am from a background in graph theory and I discovered graph databases recently. I wrote a research article with my functional point of view instead of the classical relational point of view in graph theory logic. I tried to make the introduction of my article as friendly and easy to grasp the core concepts as possible. I wanted to share my article with you : https://web.archive.org/web/20201230092310/http://lyaudet.eu/laurent/Publi/Journaux/LL2020LargeursStructuresBinaires/LL2020LargeursStructuresBinaires_v5.pdf

In the introduction I show that GraphQL queries can be solved on a reversible matrix instead of a multigraph. There is one trivial theorem where you just transform edges/arcs into new vertices using the incidency graph. And there is a slightly less trivial theorem where you transform multi-edges with properties between vertices into a single integer value from one vertex to another. The slightly less trivial theorem can, nevertheless, only be done for a fragment of GraphQL.

I don't know if any of these theorems can be made efficient and have practical applications but maybe some of you will enjoy to have a logical framework that fits graph databases.

May I ask for your feedback, please ?

Best regards, Laurent Lyaudet

[LLyaudet]

I choose feature request because I thought that maybe you would want to implement and benchmark some of my ideas.