I’m not trying to rain on any parade here…and I have no horse in the race as far as metadata is concerned (better data is good for me) but the complexity of what you are proposing is massive and while you’ve likely thought a great deal about it…I’ll just add some thoughts…as I’ve been working on/with these issues for the past 3 years in building our systems and here’s the problem…finding simplicity and a consensus by which everyone can agree…and the issue with that is you aren’t the only one working on it which then creates more confusion as to which direction to go…
Photos already have certain metadata, but photo services also store a great deal of data on the photos they represent. They have multiple identifiers for each work and sometimes these shift and change depending on regional issues. Then there are the issues regarding who actually owns what.
Then when you add music to the problem, you have multiple identification layers that extend to different owners and regions and recordings versus publishing and these often occur to describe what is essentially the same work, but perhaps different owners in different places. I mean the publishing side of this alone is likely enough to make you want to kill yourself.
So we add video to this…and text…and the complexities of licensing and so on…
But wait…then there is the fact that you speak to a decentralized/democratized system without entities that control anything and the issue here is then you come back to the problems I had to work out which includes the voracity of truth and authority. Entities actually have accepted trust and can speak truth that people believe.
So I applaud what you are attempting…again the better the data is the better our systems work, but the tech is just a first step…there needs to be an adoption plan, it would be wise to get supporters on board and also map out a timeline…and how will you fund this? It has to sustain itself somehow.
Best of luck!