1. As in the example of movie genre, it seems that some metadata still needs to be written based on a person's perspective. However, everyone has different interpretation on movies, which makes the metadata inherently commit generalization fallacy. Is there a possible way to reduce the margin of error?

2. If someone makes a meaningful analysis by collecting metadata written by other people, can the person insist on the authorship of the analysis (as artists do on the authorship of collages or ready-made objects)?

3. If metadata overturns an original meaning of data, can metadata take over the position of data?