Version control records identifiable revisions of files and their history, making it possible to inspect an earlier state and determine what changed. A version-control commit records a revision; a version-control branch permits a line of development to proceed alongside another; version-control merging combines changes and exposes conflicts requiring a decision. These operations make collaboration and investigation of software regressions practical.[ref]
Suppose a numerical simulation produced an unexpected result after a formula changed. Recording the version-control commit with the result identifies the exact source revision, so the old and new formulas can be compared and rerun with the same inputs. Reproducing the full calculation also requires the relevant data, software dependencies and execution settings: file history is one part of that record.
New to topics? Read the docs here!