not sure I understand you point ... what do you mean "if your connected graph isn’t particularly useful," who builds a graph that is not useful for something ?? 🤔
also, as I mentioned here I use a tool to build my personal documentation as a graph, so some of it is absolutely not shared understanding but I would definitely call it knowledge
It was in the context of the question raised on that thread. In terms of usefulness (which I qualified as subjective), I rather imagine most humans will, with thought, produce something consumable; machines on the other hand can construct graphs that might not be so useful. And, shared-ness is not the only qualifying criteria for knowledge; certainly not in any binary sense. Perhaps if I said it was value-adding it would be less contentious?
Véronique G. It really doesn’t have to be shared understanding. The crucial points are the relations or properties (the edges). If they transport meaning and are not simply connecting nodes to visualise them, then you have a knowledge graph. Knowledge is information+context, which you clearly have, it’s also a graph representation. Voilà.
