Identify unthrown errors in the spec.

[wip-abramson]

For example INVALID_PUBLIC_KEY, INVALID_OPTIONS.

This came about because trying to write tests for HTTP binding - w3c-ccg/did-resolution-mocha-test-suite#17

These should either be removed or a step in the algorithm that throws the error should be added

[wip-abramson]

The following errors are not referenced in the specification apart from the errors section.

INVALID_OPTIONS
One or more of the options provided for DID Resolution or DID URL dereferencing are invalid.

INTERNAL_ERROR
An unexpected error occured during DID Resolution or DID URL dereferencing.

FEATURE_NOT_SUPPORTED
The DID resolver does not support the requested feature. The value of the detail field SHOULD provide a longer description of the feature that is not supported by the resolver.

https://w3id.org/security#INVALID_PUBLIC_KEY
An invalid public key value is detected during DID Resolution or DID URL dereferencing.

https://w3id.org/security#INVALID_PUBLIC_KEY_LENGTH
The byte length of rawPublicKeyBytes did not match the expected public key length for the associated multicodecValue during DID Resolution or DID URL dereferencing.

https://w3id.org/security#INVALID_PUBLIC_KEY_TYPE
An invalid public key type was detected during DID Resolution or DID URL dereferencing.

https://w3id.org/security#UNSUPPORTED_PUBLIC_KEY_TYPE
An unsupported public key type was detected during DID Resolution or DID URL dereferencing.

Continuing a discussion we had at TPAC, if we wish to keep any of these errors in the spec then we need to identify a location in one of the algorithms where these errors are thrown.

The easiest option would just be to remove all of these errors. However my sense is that INVALID_OPTIONS, INTERNAL_ERROR and FEATURE_NOT_SUPPORTED are useful and we should identify where in the algorithms they would be thrown.

The PUBLIC_KEY errors can be removed.

Do people agree?

[w3cbot]

This was discussed during the #did meeting on 18 December 2025.

View the transcript

w3c/did-resolution#247

Wip: About errors, but not actually used. Review them, find ones not thrown and either remove or add where they belong. Many can be removed, but there are some that might be worth adding.
… Do you agree the public key ones should be deleted?

JoeAndrieu: Yes -- delete those ones as they are at a different level. Other internal ones such as options and may be raised, but at the implementation level, not in the spec algorithms.

manu: These errors came up largely through testing such as in verifiable credential testing. Errors would occur, and no error was raised -- hence these were found.
… We thought it would be good to have an inventory of likely errors, even if they weren't called out in the spec algorithms -- analogy HTTP errors.
… Feature not supported, invalid options are useful. But where we can add them to the algorithm for given errors we should add them.

Wip: Sounds good. Try to find where they fit in the algorithms, if not, try to add a generic section about other errors. Remove public key ones.

---