Which statement best describes a key differentiator of Delta Lake compared to plain Parquet files on object storage?

Delta Lake's transaction log (_delta_log) records every change to the table as a JSON or Parquet checkpoint file; enabling ACID transactions; time travel; and data versioning on top of standard Parquet data files. The incorrect options ("Delta Lake compresses data more efficiently than Parquet", "Delta Lake stores data in a proprietary binary format incompatible with Spark", "Delta Lake requires a commercial license to use in production") are distractors that don't fully capture the concept described.