Business ontology and product thesaurus: v.1 published after a year of regular iterations


Some tips to understand the v.1 of the DFC standard




May, 2018.

Since a year we’ve been working to build a common language in order to interoperate our platforms. With Bernard Chabot we went from sessions when we theorized concepts and relations, to sessions with business actors to test our theoretical modelization and evolve it step by step.

We are now reaching a point where we are confident enough in our model to test it live. So we are releasing a bunch of files that all together make the DFC interoperability standard.

Check our Github repository :-)

How we have evolved since a year on our business description

Since the v.0 we published last June, a lot of things have changed! It is really a journey to reach a point where we manage to cover all the use cases of our businesses in a simple and consistent shared model.

Here is the image of the business model we ended up with:

Products:

Transformation:

Sale session:

Transactions:

Product glossaries: the approach we choose

We studies different option for product glossaries/thesaurus. Our challenge is to make sure we can identify uniquely products from one platform to another so that information regarding this product flow without friction. As each platform has its own way of describing products, and its own logic, we choose to adopt a “facet approach” a little bit like Langual but we couldn’t use Langual as in local food system, not only food products are sold. People sell and buy soap, cosmetics, cleaning products, etc.

Also, IDs like GTIN (barcode identification) doesn’t meet our need as a GTIN doesn’t tell the precise nature of a product, like the variety of an apple. Some webshops specialized in fruits might want to classify apples in different product categories depending on variety for instance, so GTIN is not enough.

So we choose the following “criteria” to uniquely identify products:

For now we have treated physical characteristics, claims and brand as properties with only plain text content. We will probably evolve and treat them as glossaries as well in the future.

Modularization approach

For now we have only separated the business logic from the product nature logic. We want to evolve that and split the business logic to make the standard adoption easier, enable actors to adopt only one part or the other of the standard depending on their use case. Our aim is also to make the standard easier to maintain and evolve.

So we will probably go toward a standard architecture that could look like that:

We need some more thinking with Bernard and the consortium members on that, and that will be part of the work for the next iterations.

We also want to build more on “bricks” that have already been built by other actors and can solve some of our issues. For instance the “agent” ontology have been much more refinely worked on by the Virtual Assembly for some other projects. Bernard has worked with Open Source Ecology on a much more refined measure ontology that we could also use instead of rebuilding it.

Next steps

Our next step on the standard development are:

Also we are still moving forward on structuring the consortium. We are at the moment writing the “games rules” for the Data Food Consortium community and will publish them soon, using the #codesocial methodology. We are applying for some grants as well at the moment to be able to start paying the job of the team working actively on this project.

So here is where we are at! Feel free to comment on GitHub, or on this post, and contact us :-)



Share this article












Back to main website