mirror of
https://github.com/hanzo-apps/HIPs.git
synced 2026-10-30 14:47:28 +00:00
zeekay
c29a2c4a20
Deleting a proposal that is wrong or superseded is right — a document nobody should follow does not earn a place in the index by being renamed Superseded. But deletion frees the NUMBER, and the number is the part that outlives the document: it is cited in code, linked from outside, quoted in commits. Sixteen numbers were freed here today, and one was spent again within hours. HIP-0138 was published as "Cloud — Unified Hanzo Binary", deleted, then reissued for "One Manifest, Five Kinds" — so every existing reference to HIP-0138 resolved to a specification it never made. The generator could not see it: it refuses two files claiming one number, but a number claimed once and then again looks identical to a number claimed once. spent.json records what each freed number was. validate() refuses a file that claims one under a different title, and says what the number already means. The document still goes; only the address is kept, at one line each. "One Manifest, Five Kinds" moves to HIP-0142, the first genuinely unused number, with its own H1 corrected — a heading that disagrees with its front matter is how a stale number survives a renumber. docs/SCALE_STANDARD.md cited the deleted file for the per-replica capacity budget. That text lives in HIP-0122, which the lint could not know because the citation is a backticked path rather than a link. Repointed to the HIP that actually carries it, without the section number, which did not survive.