# Comments **Line Comments** start with `//`. Everything after a `//` is ignored until the end of the line. **Block (multi-line) Comments** start with `/*` and end with `*/`. Block comments can be nested. **To indicate that a comment is part of the documentation and is formatted using markdown**, use `///` for single line comments and `/** */` for block comments. ```cortex // This is a line comment /* This is a block comment */ ``` ## Documentation comments ```cortex /// This is a documentation line comment /** This is a documentation block comment */ ``` ## Comments are lossy The parser currently **discards** comments: they are not attached to the MathJSON of the following expression, and a round-trip (`parseCortex` → `serializeCortex`) does not reproduce them. Comments carry no semantic weight. The lexer recognizes the documentation-comment spellings, but the parser does not currently attach them to nodes. This is a deliberate scope decision. Notebooks keep prose in dedicated markdown cells rather than in code comments, so comment preservation is not required for the notebook workflow. (The serializer can still *emit* a `/* … */` comment when a MathJSON expression carries a `comment` metadata field, but nothing on the parse side currently populates that field.)