Open wip-abramson opened 2 years ago
Agreed on the value in a general clean up and adding an example.
As for revoke DID, I believe that is a separate function that Lohan wants to explicitly add. I'm on the fence about it.
The general term is "deactivate", which is what we have as a sequence diagram.
I believe "revoking" means actually removing the DID from existence as if its never existed. Because of the chain, it could be visible through chain analysis, but queries to the data store will just return a FAIL. I have some concerns about having these two different types of terminating a DID. It's not clear when you would use one rather than the other. IMO, the semantics of deactivate are clear, but revocation is not.
@lohanspies Can you remind me why revoke seems necessary as a distinct operation from deactivate?
@jandrieu I kept deactivate and revoke as distinct operations as @ig-shaun informed me that in Cosmos you can do either and both would be possible. My view is that we can combine them unless there is a very distinct requirement to keep both. Maybe the best approach would be to combine them and allow implementers to decide if they want to revoke vs deactivate.
Thoughts?
We should also review Heading size throughout. Currently the document has a heavily nested heading structure. e.g.
The layout feels a bit off to me currently.
It goes:
To me I read it as DID Method Name was just
did:earth
. But this title is actually defining the section where all the components of a did:earth identifier are described.I would also prefer to see an example, even if just abstract e.g.
did:earth:version:chainspace:namespace:uniqueid
, before going into each of the different sections. Suggest these might make better subheaders headers.Also, on a similar note about headings, we need to pick Revoke DID or Deactivate DID, rather than having both as headers I think