Controlling LLM Prose

Area IX · Technical Communication Formats

Semantic Line Breaks

One clause per line

Break source lines at clause boundaries so diffs and reviews stay clean.

Definition & mechanics

Semantic line breaks (also “ventilated prose”) put line breaks in the source after sentences or clauses rather than wrapping at a column width. Rendered output is unaffected, but version control shows one changed line per changed thought, making diffs, blame, and review dramatically easier on prose kept in Git.

A curiosity

Reflowed paragraphs make a one-word fix look like a ten-line change. Breaking by meaning turns your prose diffs into something as readable as code diffs.

Example

Before (wrapped at column)The client retries failed requests using exponential backoff, and it honors the Retry-After header when the server provides one.
After (semantic breaks)The client retries failed requests using exponential backoff. [newline] It honors the Retry-After header when the server provides one.

Copy-ready prompt

Copyable Prompt
Apply the concept of Semantic Line Breaks (reformat the source text with semantic line breaks—start a new line after each sentence, and optionally after independent clauses or long phrases, rather than wrapping at a fixed column width; do not change wording or rendered output, only where the source lines break) to the provided text and rewrite in place.

Apply the edits directly to the file as a clean, production-ready deliverable. Output ONLY the revised document text without embedding revision notes, changelogs, or meta-commentary in the file itself—put any explanations in your chat response only.