A technical thread on the DFINITY Forum titled “CandidX typed blocks with stable hashes under unused schema extensions” examines a narrow but important interoperability question for Internet Computer developers: how to ensure that the cryptographic hash of a typed block remains stable when a schema contains optional or unused extensions.
What the thread is about
The conversation focuses on CandidX "typed blocks" — blocks whose payloads are described by typed schemas — and the behavior of their serialized form when a schema includes extensions that are present in the definition but not used by a particular block instance. The core question is whether and how those unused extensions should influence the block’s serialized bytes and therefore its hash.
Why this matters (in plain language)
A block hash is a compact fingerprint of the block’s serialized contents. If two nodes serialize conceptually identical blocks differently, they will compute different hashes. That matters for any systems that rely on stable, deterministic hashes for verification, indexing, or cross-node comparison (for example, client tooling, indexing services, or cross-system verification logic).
Key technical themes in the discussion
-
Deterministic serialization: Participants examine the need for a canonical encoding so that the same logical block always yields the same byte sequence, regardless of whether optional schema fields exist but are unused.
-
Treatment of unused schema extensions: The thread considers whether unused extensions should be omitted from serialization, serialized in a fixed placeholder form, or handled by an explicit rule in the schema/serialization layer.
-
Schema evolution and compatibility: Ensuring stable hashes while allowing schemas to evolve (add optional fields/extensions) is a recurring concern; approaches that preserve backward compatibility without producing hash churn are of interest.
-
Practical developer impact: The topic is relevant to developers building canisters, indexing services, block explorers, and other tooling that reads or verifies blocks and depends on consistent hashing behavior.
Status and next steps
The conversation is a technical forum discussion and does not represent an implemented protocol change or a governance decision. It is part of ongoing developer and protocol-level conversation about serialization, hashing semantics, and schema handling on the Internet Computer. Developers and operators who rely on deterministic block hashes may want to follow the thread for proposed approaches, trade-offs, and any future formal recommendations.
If you work on tooling or canister code that depends on block hashes or on Candid/CandidX schemas, this thread provides detailed developer-level considerations about how unused schema extensions interact with hashing and serialization.