Abstract
Our information systems are siloed and based on heterogenous labels, which
breaks real interoperability, automation or machine learning algorithms use.
We need to exchange information between humans and machines to inter-operate
from multiple data domains sources, across various actors, tools, organization
branches, internally with our sites and affiliates and externally with our
multiple partners.
Following ISO15926 part14 principles, published by PCA this summer, associated
with other standards of primary interests such ISO/IEC81346 eg, lessons learnt,
and experiences acquired with AFNOR IDMI[2], IOGP DISC[3], JIP33 & JIP36
CFIHOS[4] and others working groups, we started in January 2021 to implement a
semantic framework to set-up a governance methodology, with rules, common
principles, and guidance's to provide tools to sustainably structure our
information legacy in a normalized way.
By bridging our data and documents texts to selected and combined standards
classes, using W3C standards, respecting the same context, exclusively
structured through ISO15926-14 properties to set the relationships between
classes, we allow working as a continuum between conceptual, logical, and
physical models using trustable labels that recovered their sense.
To govern this common framework, we developed a mapping strategy based on
several steps to enforce and endorse progressively our company reference
ontology as a result of our mapping efforts to link our data to standards; this
disambiguation or contextualization process loop is a way to formalize our
verifiable requirements thru a generic specification enabling data exchange, in
a way (or a model or a specification) which is not specific to TotalEnergies.
This approach is an intellectual and cultural breakthrough, beyond our current
information management way of working; it implies a change of mindset, supports
a progressive participation of the stakeholders for connecting our data and
taxonomies to higher level conceptual common nodes. We believe this a one of the
keys to debottleneck our fragmented information system.
To support adoption by our organizations, we shall focus attention on pedagogy
while each Use Case's resolved mappings are sustainable, means qualified and
re-used by third POC or business. Finally, we do believe that a common language
across our disciplines and métiers is key, this standard language creates
unambiguous understanding among all stakeholders internally and externally. This
is a matter of education, federated and shared practices, and adapted tools to
support the methodology. It provides user experience customized interfaces by a
modular and flexible approach, and the development at scale of domains and UCs
knowledge graphs to explore, navigate and use FAIR[5] and safe data.
The relation of this journey in a corporate conducting its digital transition
will be illustrated by concrete outcomes of this initiative and will give firsts
lessons learnt up to now, as ground of further requirements.
[1] SMART: Specific, Measurable, Achievable, Realistic, and Timely
[2] IDMI: Ingénierie des Données et des Modèles pour l'Industrie
[3] DISC: Digital Information Standards Committee
[4] CFIHOS: Capital Facilities Hand Over Specification (IOGP JIP36)
[5] FAIR: Findable, Accessible, Interoperable, Reusable
#### Auteurs/Autrices
Linkedin [Jean-Charles Leclerc](https://fr.linkedin.com/in/jean-charles-leclerc-54bbb586)