This is a huge step in the right direction. Definitely, reputation is multi-dimensional, and I love the fact that you haven’t limited the number of dimensions someone can report. Self-reporting is problematic for many reasons, but right now this is the most sophisticated system I’ve seen. I imagine in the future my phone/computer will auto-record what I am actually doing, and if the machine learning has enough tags, the system will be able to self-label as well, allowing me to give feedback if I want things in a different category (within reason, as in, if I’m playing a game, it won’t let me say I was writing a document, but it would let me say I was talking to my mom as I was playing the game. Not saying my son does that, but just as an example.)
One of the most interesting next steps would come when the reporting is based on actual behavior, and the results of the behavior. For example, if I wrote a whitepaper (I noticed you did that), and the next person made minor corrections (or paid me for it), that’s different than if I wrote a whitepaper, and the next person had to rewrite the whole thing.
Another important example is in voting and decision-making. I don’t visibly a lot of time voting or decision-making, but some of my decisions are better than others. If we are talking about a voting system, and my reputation in a particular realm, you would want to know if my past decisions in that specific realm are wise or not.
A third area is the area of knowledge. If I am voting in a particular domain, how much information did I consume in that domain, and how many different perspectives did I consume? I heard recently about a voting system where a person needs to see 3 different perspectives and pass a mini-test that they understand the issue before they get voting rights in that community. I’m not saying that’s the best way, but it’s one way to consider what reputation is.
All in all, what you have built here is extremely impressive and a huge leap beyond any reputation system I’ve seen so far.