timestamper with null timestamper option

README.md

···
 
 Other parts of Atproto support two address standards, DID:Web and DID:PLC. DID:Web is unsuitable because we need to maintain a history of updates which cannot be tampered with even by the timestamper themselves. Using DID:PLC would add a degree of recursiveness to the system which makes it confusing to analyse.
 
-Timestamper key rotations are expected be unusual so the drawbacks of a public blockchain (cost per transaction, slow confirmation speed) should not be disqualifying. We propose DID:ETHR because it already exists as a well-specified standard (albeit not a widely-adopted one). Using a simpler bespoke standard instead would also be viable.
+Timestamper key rotations are expected be unusual so the drawbacks of a public blockchain (cost per transaction, slow confirmation speed) should not be disqualifying. We propose DID:ETHR because it already exists as a well-specified standard (albeit not a widely-adopted one). Using a simpler bespoke standard instead would also be viable. Another option would be to use DID:PLC, but with the additional restriction that the `timestamperDid` of a timestamper must be null so all updates to it would have to follow the "forced update" path.