How to distinguish between witness proofs and controller proofs

[brianorwhatever]

Currently the only difference between a witness proof and a controller's proof is that the witness will likely be a did:tdw. This isn't a particularly useful way to determine which is which. I think we need a better technique.. might be one reason for #66

[swcurran]

As I was writing this — I think the answer is that the DID of the DID Controller MUST be a did:key, and MUST be in the authorized key array. Conversely, witness did’s MUST NOT be did:key, and witness DIDs MUST be in the witness list.

A bit of data sniffing, but I don’t think outrageous. The question is whether it is better or worse than using Proof Chains, given their lack of implementations?

[swcurran]

@brianorwhatever -- do you want to see this added to the spec? Seems it would be useful.