DocCommuns - France
= "DocCommuns (*produced with: Coexiscience): Building a knowledge base on commons based on the work of La Coop des Communs": the documentary database of the Coop des Communs"
URL = https://nextcloud.coopdescommuns.org/index.php/s/Hr7exneiRWicNG6
Description
"La Coop des Communs now has an analysis of the documentary productions of its working groups (books, research papers, evaluation reports) covering the past 9 years. This is the first result produced by a prototype developed by Coexiscience, with funding provided by La Coop des Communs in 2025, using AI, particularly LLMs (Large Language Models, ChatGPT and others).
This prototype is called DocCommuns. It is a knowledge base that we have begun building with Coexiscience, by facilitating interaction between AI and humans, between AI researchers and "experts" on a given subject (in this case: the commons) (NB: a knowledge base is an organized information storage space that can be queried by combining organized knowledge).
DocCommuns has thus allowed us to analyze the 27 documents produced by the working groups of La Coop des Communs since 2018, according to a set of criteria. This framework was designed based on input from a group within La Coop des Communs that met in late 2023/early 2024, with the goal that our association should make public policy proposals in favor of the commons. Easier said than done if you want to be rigorous. The work of designing the analysis framework and writing the criteria was carried out by Nicole Alix between summer 2024 and summer 2025. "
Characteristics
"It resulted in 63 criteria grouped into 3 chapters:
Objective A: Foster the collective self-organization of civil society groups
Objective B: Act for the financial autonomy of the commons, adapt funding and evaluation to the specificities of the commons
Objective C: Obtain institutional recognition for the commons. "
How To
"The DocCommuns knowledge base makes it possible to answer queries (questions or requests for information) such as: "which documents address criteria A5 and C7 but not B3?" or "which criteria are centrally addressed in documents D12 or D23 and are addressed directly or centrally in D4?" So, for each of the analysis criteria, the analysis carried out with DocCommuns allows us to trace, document by document produced by La Coop des Communs, how the issue is treated, in what manner, with which words, what key lessons are drawn, on which page of the document, and what examples are cited.
The result of our work is now accessible via a summary document of all the analyses carried out by the process: 600 pages in total."