Should we add identifier persistence to the identifier spec

[zachzeus]

[### Impacted sections ]

https://uncefact.github.io/spec-untp/docs/specification/DigitalProductPassport ](https://uncefact.github.io/spec-untp/docs/specification/IdentityResolver

Issue Description

In the discussion on 7/Feb we discussed identifier persistence - there was concern that this might veer into legislative side.

Maybe we should add a SHOULD requirement that tells people that identifiers should be persistent, but doesn't require it.

Remember to add relevant labels.
]

[Tweeddalex]

Hey @zachzeus, I'm aligned with this in principle; however, in practice there are only a few DID method candidates that could handle persistent identifiers. The few that spring to mind would be:

Hybrid DID methods

• did:webvh (formerly did:tdw)
• did:webs (although you could argue that KERI on its own isn't persistent)
• did:scid (still up and coming)

Blockchain based DID methods

• did:cheqd
• did:ebsi
• did:iota, did:iden3 (not as mature)

For this reason, I think it is something we should continue to ask ourselves - as there is good work going on to get to this point. Additionally, if we did want to go with a persistent identifier type, then we'd also need to consider the governance/security/legal requirements for all parties involved (which gets a bit tricky!)