Compare file content without trusting names or dates

Two copies can have different names but identical contents, or familiar names but different bytes. Compare that evidence explicitly when colleagues need to identify a shared artifact. This manual exercise uses harmless invented local text and Python's SHA-256 calculation. A Kiguri office can host the discussion; no built-in file verification, personal export, or upload service is assumed.

Define the bytes before using the names

Create two practice values with exactly the same ASCII content: Guide v1 followed by a line-feed character. Call them draft.txt and renamed-copy.txt in the comparison record. Those labels are illustrative; the fixture check can operate on byte values in memory without reading private files.

Each value contains nine bytes. A third fixture contains Guide v1 followed by carriage return and line feed, giving ten bytes. The displayed words may look the same in a text viewer, but the chosen byte representation differs. Record encoding and newline assumptions so another reviewer can reproduce the comparison.

Compute the same digest method

Using Python hashlib.sha256 on the nine-byte value produces a75521d8a1aa9a56 4ebbf8b4492835ab 1c18d2042849f0bf 59065723d4fa652e. The digests here are displayed in four groups of sixteen hexadecimal characters; remove the display spaces when comparing the full digest. Both identical fixtures produce that digest, despite their different illustrative filenames.

The ten-byte carriage-return/line-feed fixture produces 78bbd03c1229b538 2420f22f6fa3074a cf8fdf1287997192 b7aef9d85b9ee006. Keep the method, byte length, and digest together rather than comparing unexplained strings. Compute on the actual bytes intended for comparison, not on a viewer's reformatted text or a filename. These exact fixture outputs can be checked locally without contacting a server.

Separate a mismatch from its meaning

Here the different digest reflects a known newline difference. It does not establish a different guide instruction. Conversely, identical fixture digests do not tell the group whether Guide v1 is approved, current, correct, or written by the expected author.

Inspect the representation behind a mismatch before calling a copy damaged. A format can contain metadata or use different serialization while presenting the same visible content. That question needs an appropriate separate comparison; a byte check does not establish visual or semantic equivalence. Preserve the original values rather than silently normalizing them and reporting the normalized result as the original evidence.

Keep authority separate from content evidence

Attach the comparison record to the team's agreed authoritative artifact location. Record which copy the responsible owner identifies as current. Do not let a matching digest substitute for that decision, or treat a newer filename as evidence that the content changed.

For actual permitted files, follow the team's normal handling rules and keep sensitive material out of demonstration records. This exercise needs only invented bytes. It makes no security certification, authorship assurance, or promise that a particular file is safe to open.

Ask what the comparison established

Have another participant explain why the first two fixture names do not imply different content, and why the third fixture can differ at the byte level without different displayed words. They should also identify what remains unknown about correctness and approval.

The useful artifact is a local comparison record with method, input assumptions, lengths, digests, and scope. It supports an evidence-based copy discussion while leaving meaning and authority to the checks that can actually establish them.