Open pfeairheller opened 1 year ago
So we will support three did methods?
I think there should always be a did:keri because its recognized now and will gain more recognition over time. But having a more secure version of did:web that is a limited version of did:keri makes some sense also. Not sure.
Yes, that makes more sense to me. All three methods, we are defining the last 2 here. did:keri includes the did:webs approach but also the super watcher variant. did:webs is just the variant with the domain name.
So maybe we need a did:webs repo and all this initial work needs to go there.
And if we are really successful did:webs will supplant did:web
I'd like to see all of following variants be possible, is that consistent with what you're saying above @pfeairheller ?
Variant 1:
did:keri:<aid>?oobi=
-> DID URL gets dereferenced to DID document using "oobi" (or similar) parameter
Variant 2:
did:keri:<aid>
-> DID gets resolved to DID document using OOBI(s) or super watcher that are either preconfigured in the resolver or passed to the resolver as options
Variant 3:
did:web(s):<domain name>:<path components>:<aid>
-> Resolved just like did:web with additional KERI verification logic happening
A useful analogy. did:keri is to did:web what https is to http. And there should be about the same level of effort to step up from did:web to did:keri. Let's people start with the easy way, but people will quickly transform to using the secure way because they remember http/https.
Instead of did:keri, we call it did:webs and push that analogy. From a marketing perspective it’s better - not named for KERI, but powered by it. Clean slate in some circles.