I have always felt a bit uneasy to say that I am working in “Knowledge Management”. “Knowledge” is such a big word, that I always wonder what my interlocutor (or reader) think “Knowledge Management” is.
So I found myself recalling a conversation going like: “I am building a Knowledge Base” and receiving as an answer: “so you are building a database?”. The difference is intuitively obvious to me, but how can I simply explain it?
Of course I could present many definitions of “Knowledge base” and “Database” to clarify the difference. I could go through the data, information, knowledge, wisdom pyramid, or I could refer to Knowledge bases as expert systems. I could consider the distinction between implementation and logic layers.
But in the face of such a simple question, I think all these definitions would be too elaborated, and unconvincing.
So I have been trying to formulate a very simple explanation, and the best I came to is to rely on my intuitive distinction of databases and knowledge-bases, with a simple example:
Imagine we are talking about food recipes.
When I build a database, I would focus on a model to represent ingredient, processes, recipes and all elements that are part of this domain, with the intention to fill it.
When I build a knowledge base, I would assume a that database exists, or start with a significative first amount of data. Then I would be looking at enriching it: what taxonomies/ontologies can be used to aggregate ingredients? can I enrich names with synonyms from other contexts or languages? Can I link meals, recipes, ingredient to other information (nutritional databases, chemical databases, Wikidata…)? Eventually I would have some classes of operations/questions in mind to guide my enrichment, and in some cases add some logic or other form of inference to automatically derive additional information from my data.
They are really different things to do. One is building a structure to collect data (and eventually collecting it). The other is analysing many resources and inferring relations.
