if prev and next links point to bare hashes, they can be assumed to be updates at the same domain (or domain+path). but if they are full DID URIs, i.e. did:webhash:old-example.com:{genesishash} or did:webhash:new-example.com:{old-genesishash}, would they be valid?
Alternative is to say that migrations are only recognized unidirectionally (i.e. with a next, not bidirection with a prev ), and make the first did doc at any given domain or domain+path a new genesis hash and new entropic identifier. This avoids the breakage if a prev link points to a dead site or a site that deleted old did docs, but creates more complexity for the witness/indexer services that track users across migrations
if prev and next links point to bare hashes, they can be assumed to be updates at the same domain (or domain+path). but if they are full DID URIs, i.e. did:webhash:
old-example.com:{genesishash} or did:webhash:new-example.com:{old-genesishash}, would they be valid?Alternative is to say that migrations are only recognized unidirectionally (i.e. with a next, not bidirection with a prev ), and make the first did doc at any given domain or domain+path a new genesis hash and new entropic identifier. This avoids the breakage if a prev link points to a dead site or a site that deleted old did docs, but creates more complexity for the witness/indexer services that track users across migrations