diffs

Diffs are representations of the differences between two sequences, most commonly text files. A diff describes the changes needed to transform an old version into a new one, typically as additions, deletions, and modifications of lines. They are used to review changes, publish patches, and support version control workflows.

Common diff formats include unified diff, context diff, and the traditional normal diff. Unified diffs show surrounding context and annotate added lines with a plus sign and removed lines with a minus sign, making them easy to read for humans and for tools that apply patches.

Diffs are generated by comparing the two inputs and identifying a minimal set of edits. Algorithms such as Myers' diff algorithm are widely used to compute diffs efficiently, especially for large text files. The output can be used directly by patch utilities or by version control systems to display changes.

Version control systems such as Git, Subversion, and Mercurial produce and consume diffs to show changes between commits or working trees. Patches, which are diffs packaged for distribution, can be applied to another copy of the repository or sent for review in collaborative workflows.

Limitations include that diffs are usually line-oriented and may not reflect intent or logical changes in refactoring; large rewrites can produce large, hard-to-interpret diffs; binary files require different handling or a non-text comparison.

Diffs date from early Unix tooling; they remain a core concept in software development, enabling incremental collaboration and change tracking.