dataHash CIP

datahashlasmlabscity.md

@@ -0,0 +1,96 @@
+---
+CIP: ?
+Title: New Plutus built-in `dataHash`
+Authors: Las Safin <las@mlabs.city>
+Discussions-To:
+Comments-URI:
+Status: Draft
+Type: Standards Track
+Created: 2022-02-15
+* License: Apache-2.0
+* Requires: 35
+* Replaces: 36
+---
+
+# New Plutus built-in `dataHash`
+
+## Abstract
+
+We propose adding a new built-in `dataHash` for calculating the hash of `Data` on-chain,
+in such a way that it is coherent with `txInfoData`.
+
+## Motivation
+
+This is useful for essentially "putting" `Data` into token names, the contents of which a script
+can access by having the concrete `Data` be passed in the redeemer (whether it's a minting policy
+or validator).
+
+It is also possible to verify signatures of (hashes of) `Data` with this built-in.
+
+This likely has many other uses, as it's a powerful feature.
+
+Currently it is essentially infeasible to do this without the built-in, as it's
+very expensive to first serialise the `Data` into CBOR then hashing it using `blake2b_256`.
+
+## Specification
+
+### Function definition
+
+The new built-in is:
+
+```haskell
+dataHash :: data -> bytestring
+```
+
+The reason for the name is that an off-chain `dataHash` already exists.
+
+The semantics should be such that for any key-value pair `(k, v)` in `txInfoData`,
+it should be true that `k == dataHash v`.
+
+### Cost model
+
+The CPU cost should be linear in the size metric of the input.
+The memory cost should be constant, possibly the exact same
+as for `blake2b_256`.
+
+## Rationale
+
+A possible alternative is CIP-36, which proposes a function `serialiseData`, such that
+`blake2b_256 . serialiseData` is equal to `dataHash`.
+
+The reason `dataHash` is considered to be a better approach by the author,
+is that in practice, all sensible uses of `serialiseData` are also possible
+using `dataHash`.
+
+Hashing using a different algorithm isn't something you need in practice,
+because no existing protocol hashes Plutus's `Data` with something other
+than `blake2b_256`. If in a future version of Plutus the hashing algorithm
+were to be changed, then `dataHash` would also be changed accordingly.
+
+With this approach, you can not inspect the raw serialised `Data`,
+however, this is not something that is useful in practice, as the
+serialised `Data` doesn't give you any useful information other
+than its hash.
+
+Another important point to note is that `dataHash` is potentially more
+efficient on large `Data`, since the hash can be computed without storing
+the entire serialisation of the `Data` in memory.
+
+Implementation-wise, it should also be simpler than `serialiseData`, since
+there are no considerations wrt. what the format of the serialisation should be,
+because it's never exposed to users.
+
+## Backward Compatibility
+
+Adding this built-in doesn't break older transactions. There is the option
+of adding this built-in in only Plutus V2.
+
+## Implementations
+
+Issues for the status of the Implementation:
+- https://github.com/input-output-hk/plutus/issues/4167
+- https://input-output.atlassian.net/browse/OBT-370
+
+## Copyright
+
+This CIP is licensed under Apache-2.0