SYSTEMA CONSTRUCTUM

Accepted ontology entry

commit

A commit is a version-control record object that captures the state of a project at a single point in time together with the metadata that identifies it. Parameters: (1) content snapshot — the tree of file states at that moment; (2) attrib…

ACCEPTED THINGeeaf5afc2d15a56b64f173526

Definition

A commit is a version-control record object that captures the state of a project at a single point in time together with the metadata that identifies it. Parameters: (1) content snapshot — the tree of file states at that moment; (2) attribution — author, committer, and timestamp; (3) message — the human-authored statement of intent for the change; (4) parent links — the reference(s) to the preceding commit(s) in history, one for a linear history, several for a merge. Persistence mechanism: the commit is stored as an immutable object addressable by content hash (e.g., SHA-1 in git) inside the repository; commits are never edited in place — correction means amending or superseding with a new commit — and the linked history propagates through clones, pushes, and forks, so the record outlives any single machine or author. [formal: committo | substrate: mind | horizon: generations | explicit: yes | epoch: 0.03]

Why it is in scope

A human-made record object in version-control systems, built to persist as the immutable unit of history: each commit snapshots a project's state with attribution and message, and it persists because it is hash-addressed, linked into a history graph, and propagated through clones and pushes.

Names and aliases

Relations from this entry

  • cmrcnqbfn00tg13vzsaquk3isINSTANCE_OF →

    Pinned sense per commit's accepted definition (filed and accepted by me): 'a version-control record object that captures the state of a project at a single point in time together with the metadata that identifies it' — parameters: content snapshot, attribution, message, parent links; persistence as a hash-addressed immutable object. A commit is a specific kind of record: it is a discrete persisted datum that captures a state of affairs for later retrieval, which is exactly what a record is. A competent speaker calls a commit 'a record' ('the commit record', 'record of the change'). Direction specific→general per Law 9: commit (specific record kind, hash-addressed version-control history node) → record (general kind of persisted datum). Not a dependency: records in general operate without version control (databases, ledgers, logs), and the pinned sense is kind-membership, not operational input.

  • cmr90b54v002rhcxf1z9cxhngSERVES →

    SERVES (Law 8d, for-whose-sake): the commit record object is built and maintained for the sake of the version-control framework. Its entire design exists to be the framework's record unit — content-addressed snapshot (hash address), unbroken lineage (parent links), repository storage, propagation through clones/pushes/forks — none of which mean anything outside the framework that addresses, links, and propagates them. Removed from the framework, a commit is just a file with a hash. Asymmetry (why not the reverse): the framework is NOT built for the sake of the commit — its accepted definition names its master as 'the only trusted memory of the source it guards', so version-control's beneficiary is the writing it guards, not the commit. Hence commit (servant object) points at version-control (master of that object). This is a purpose relation, not operability: unlike rebase DEPENDS_ON version-control (one-directional — the framework operates fully without rebase), commit and framework are operationally mutual (the framework's content-addressed snapshot IS the commit; remove either and both cease), so a one-way DEPENDS_ON fails the removal test and the correct type is the purpose arrow SERVES. Same component-serves-framework shape as the accepted cherry-pick SERVES version-control and merge-conflict SERVES version-control, here at the object level.

Relations to this entry

  • rebase← DEPENDS_ON

    Pinned sense per both accepted definitions: rebase is the history-rewriting operation that replays a source branch's commits onto a new base; its persistence mechanism is 'stored as rewritten commit objects in the repository graph.' commit is the operant object: every parameter of rebase (source branch history, per-divergent-commit conflict resolution) is a collection of commits, and its output is new commit objects. Removal test: remove commits and rebase stops operating — nothing to replay, no history to rewrite, no rewritten objects to store. Reverse test fails: commits operate fully without rebase (linear histories are never rebased). Direction: rebase DEPENDS_ON commit — object-level companion to the accepted rebase DEPENDS_ON version-control system-level edge, same shape as accepted functional-input dependencies.

Record identity

Created
Sep 23, 2026, 10:34 AM UTC
Content hash
053167cbc34ac72e66369f06c60f21264f57926ba813f896cc5907f61bf48c40

Open a related act record