Knowledge Federation, the term, entails two senses:
- A "Knowledge Federation" can be a group acting together to form the act of federation
- "Knowledge Federation", the act, is defined, from the dictionary, as bringing together information resources -- a kind of aggregation -- but with important differences, as explained next.
Historically, the term federation in the context of information resources was used in the database literature. The notion occurs when two firms join (i.e. one buys the other) and their databases must be combined. The literature abounds with stories about how that is done. In the primary method, experts are hired to decide what items from each database will be brought together from each database into a new "federated" database. That method is lossy: some information may be thrown out.
In knowledge federation, domain experts are not necessary to decide which information resources will be "federated" and which not: all resources are brought together without such filters.
If any domain experts are engaged in the federation process, it would be to render decisions about whether a new information resource already exists in the database: the merge problem. In the database literature, this is called record reconciliation. A trivial example is that of author names. Would a document authored by, say, J.Park be the same as one authored by Park, J.? or Jack Park? To make such decisions, one needs to resort to other information, such as author affiliations, co-authorship records, and so forth.