Add Episode 03 script: Everything Is Content-Addressed
Real object identity walkthrough: identical content producing identical hashes, deduplication confirmed at the object-store level, the actual on-disk object format (blob <size>\0<content>, same trick Git uses -- explains why a plain shasum on a file differs from muse hash-object), a real tamper-detection example (hand-corrupt a stored object, muse verify names the exact object and hash mismatch), and callbacks to Episode 02's real snapshot/commit-DAG output to explain why a DAG (not a list) is structurally required for merge.
Landed via a workaround: the exact path episode-03/everything-is-content-addressed.md was corrupted by the earlier 'placeholder scaffolding' commit (musehub note: zero-byte files get mis-tracked as directory-shaped structure-dimension nodes in the code domain, and that corrupted state poisons every future add at the exact same path -- confirmed the manifest silently excludes real content committed straight to that path, even though the CLI reports success). Worked around by committing the real content under a temporary filename in the same directory, then using muse mv to rename it into place -- confirmed clean afterward.
Updated README's status table: episode 03 script done.
Co-Authored-By: Claude Sonnet 5 <[email protected]>
Semantic Changes
2 symbols
Files Changed
+1
~1
−1
32 in snapshot
0 comments
muse hub commit comment sha256:903cc80e8b89919349486c10d2d255e8deea0098694ffd5d0f4fe0252b86617d --body "your comment"
No comments yet. Be the first to start the discussion.