Information can vary over time, come from many internal and external sources, and be based on different identifiers and models. These multiple versions of the truth (MVOT) need to be reconciled on access by context.
Different sources might not always be consistent about the same Data Point. And that may be legitimate for various business, geographical, privacy, legal or timing reasons. Rather than try to force everything into a single overruling set of facts, an EKG allows them to coexist, with the access context used to make coherent selections which are consistent within themselves. In order to make those selections multiple people and systems must have transparent access to all facts (including source, identity, meaning and value-at-time information) about all objects. Machine-executable business rules may be used at query time to join instances of data and establish quality rankings.
This approach allows the EKG to represent the sometimes messy reality, which encompasses not only a variety of different organizations and systems within the enterprise, but also external sources.
Attention needs to be paid to maintaining sufficient context with each data set, and considering what is needed for each data usage use case or context, including quality.