Ordinals inscriptions

Inscriptions embed arbitrary data (images, text, code) in the witness portion of taproot script-path spends, wrapped in an "envelope" of no-op script opcodes. Ordinal theory is an off-chain accounting convention that numbers individual sats by mining order and tracks them first-in-first-out through transactions, letting an inscription be treated as attached to a specific sat. Two structural facts: the data rides in consensus-valid transactions (nodes cannot distinguish "spam" from "legitimate" by consensus rule), and witness bytes count 1 weight unit versus 4 for other bytes (see block-size-weight), so bulk data pays roughly a quarter of the fee rate of equivalent non-witness bytes.

Nuance: Sat numbering and inscription ownership exist only in ordinals-aware indexers — base-layer consensus knows nothing of them. Arbitrary-data embedding predates ordinals (OP_RETURN, fake pubkeys); taproot's removal of script-size limits made large payloads practical. Whether node policy should filter such transactions is a live policy debate (claims territory), distinct from these mechanics.

Common misstatements:

• "Ordinals broke Bitcoin's rules / exploited a bug." — Inscriptions are consensus-valid uses of taproot script; the debate is about policy and fee-market effects, not validity.
• "NFTs are stored on Bitcoin like on other chains." — The data is on-chain, but the "token" semantics (which sat, who owns it) are an off-chain indexing convention.
• "The witness discount was created for inscriptions." — It shipped with SegWit in 2017 to price witness data's lower long-term cost; inscriptions arrived in 2023.