We could have setup a standardization group within GS1. We chose not to as GS1 were asking us to sign a form to tell we commit to use the GS1 standards. How can we commit to use a specific solution when we have not yet investigated our problem, the potential solutions available, etc. ? We are pretty happy that we made this decision, as we remained free to work with Open Food Facts for products identification, which wouldn’t have been possible in the context of a GS1 standardization group.
We need a general common universal product id base to be able to match products between platform, and know if we talk about the same product, or another. In the agro-industry and retail industry agents use the EAN, GS1 product id codes. In the context of local and alternative food systems, short food chains, very few producers have subscribed to GS1 services, they don’t have EAN. And probably most won’t want to subscribe and pay the 85€ min yearly fee. So we have chosen to collaborate with Open Food Facts and use their identification system as our universal products identifiers. When there is an EAN, Open Food Facts use it as product identifier. When there is not they attribute one randomly, for free. We can call those ids “pseudo EAN”.
As we don’t want to force all actors to open their data (this should remain the agent decision), we will have our own “DFC universal product” catalog but we will use Open Food Facts API to generate ids for products which don’t have EAN. We also have chosen to use the same taxonomies as Open Food Facts to describe products, so a product file in the DFC catalog will be super easy to open and duplicate into Open Food Facts anytime the agent want to open their data.
We also chose a “neutral” agent to host the shared data of the consortium: The DFC universal products catalog. This avoid any conflict of interest amongst the consortium partners.
When building both the semantic and technical specifications of the standard, we need real life cases to test whether what we imagined is adapted to our needs and facilitates our work. So we are developing a prototype not only to create a “proof of concept” and show the potential of the standard to solve the problems mentioned above, but also to build and improve step by step the semantic and technical standard.