i've designed a nostr-based DNS (and TLS) replacement and it requires some kind of consensus to function, it uses an explicit trust graph created by the name server nodes (i'm calling it FIND, Free Internet Name Daemon) to evaluate name registration/transfer proposals and store their evaluation, anything that falls under 50% is not stored by a node. this means that the consistency is not strong, but weak, but it is weak eventual consistency, and instead of using a monetary value incentive by being a currency, it provides a directory of the names for anyone who hosts a replica to use in their apps, and secondly, the names themselves become an ad-hoc non-fungible asset (so for which reason there is a number of measures to stop perma-registering names, and other spam mitigation methods).

idk how Kaspa can be not a blockchain, unless you are splitting hairs by distinguishing a blockchain from a tangle or a DAG type structure. they are still fundamentally the same thing, blocks of data that have back-references that build a DAG. chains just aim to grow one branch where tangles can vary from one to ten or more concurrent and interlinked chains. they still need some kind of spam limit mechanism to control how many tokens and transactions are processed and all that, and are based on a token ledger.

what i have designed is not a token ledger it's a registry for unique symbols and it doesn't need that economic basis to justify its existence. it's useful in itself as a directory cache, and useful for people in that these names can help increase the appeal of an offering. it's not just a collectable, and it's not fungible, it is an actually useful thing "this name was registered by this npub".