CORE — Celestial Object
Renaming Engine

1. The Scale Problem

The International Astronomical Union has officially named fewer than 450 stars. Every other star, galaxy, pulsar, and black hole in every catalog humanity has ever compiled carries only a catalog designation — HD 189733, NGC 4889, PSR J0437−4715.

The Preservation Network holds:

The engineering challenge: generate 55 million or more names that are (a) globally unique, (b) pronounceable in any human language, (c) grand and memorable, and (d) deterministic — the same object always receives the same name regardless of when or where the engine runs.

2. Naming Philosophy

Multi-cultural roots

Historical star names draw on Arabic (Aldebaran — "the follower"), Greek (Arcturus — "bear guardian"), Latin (Vega — "swooping eagle"), and Sanskrit. CORE names must feel like they belong in this tradition without borrowing directly from it.

Root material is drawn from phonological patterns across Arabic, Sanskrit, Swahili, Nahuatl, Polynesian, Norse, Latin, and Greek — giving names a cross-cultural acoustic resonance rather than a single national sound.

The rule: no borrowed star prefixes

A name cannot begin with the same phoneme sequence as an existing named star. Regulina from Regulus, Altarina from Altair, Polarisa from Polaris — all forbidden. The names must stand alone.

Grand and light

The target register is the one real star names occupy: three to four syllables, open vowels, liquid consonants (L, R, M, N), a sense of weight without heaviness. Not Japanese. Not Latin scholarly text. Not computer-generated syllable strings.

3. The Design Journey — Five Iterations

Each iteration below solved the previous iteration's primary failure. The full sequence is documented here so future engineers understand why the final design is shaped the way it is.

4. Final Architecture

Positions 1, 2, and 4 draw from the same prefix-free pool of 422 curated word roots. Position 3 draws from a fixed set of 11 binding phonemes. Total name length is 7–11 characters, averaging approximately 9–10.

5. The Binding Phoneme — Why It Works

Real star names, examined phonetically, reveal a recurring structural pattern: a heavier onset, a secondary syllable, a light open phoneme, and a closing syllable. The light phoneme in the middle — almost always a single vowel — is what makes the name flow rather than march.

The 11 binding phonemes

Position 3 allows five pure vowels and six classical diphthongs:

6. Name Space Mathematics

Total unique names = P₁ × P₂ × P₃ × P₄, where P₁ = P₂ = P₄ = root pool size and P₃ = 11.

Prefix-free constraint

Positions 1, 2, and 4 draw from the same pool. For the bijection to hold — no two distinct (P1, P2, P3, P4) tuples producing the same name string — no pool entry may be a prefix of another. If ve (2-char) is in the pool, then vel, ven, ver (3-char entries starting with ve) are excluded. Validated at engine startup.

Bijection — the uniqueness guarantee

A fixed seed (0xA57E4321) generates four permutation arrays over the pool. Each object's global rank maps through these permutations to exactly one (P1, P2, P3, P4) tuple, producing exactly one name. A scatter multiplier is applied first so that consecutive ranks produce phonetically diverse names — not a run of names with the same prefix.

# Core bijection — simplified
def id_to_name(global_rank):
    scattered = (MIX_A * global_rank) % TOTAL     # spread consecutive ranks
    idx = scattered

    p4 = idx % N;   idx //= N
    p3 = idx % 11;  idx //= 11                    # binding phoneme position
    p2 = idx % N;   idx //= N
    p1 = idx % N

    name = POOL[P1[p1]] + POOL[P2[p2]] + BINDING[P3[p3]] + POOL[P4[p4]]
    return name[0].upper() + name[1:]

7. The Root Pool

Roots are curated — not generated — from phonological material across 25 letter categories. Each category contributes approximately 12–20 roots: 2-char roots (usually 3 per consonant, vowels a/i/o) and 3-char roots for the remaining vowel slots.

Root categories

8. Global Namespace — What Each Type Gets

Stars, galaxies, pulsars, and black holes live in separate database tables with separate ID sequences. To guarantee that no two objects of any type ever share a name, CORE assigns each type a global offset: a reserved, non-overlapping slice of the 826-million-name space.

Stars receive the overwhelming majority

Stars get 700 million slots — 84.7% of the entire name space. This is deliberate. CORE names Milky Way stars only. The stars of other galaxies number in the hundreds of trillions; even if every one were catalogued, naming them individually would be an undertaking of an entirely different order. The Milky Way itself contains an estimated 200–400 billion stars. 700 million slots covers the full realistic scope of future Gaia-era surveys with room to spare.

Each object's global rank = its offset + its sorted position within its type. This rank feeds directly into the bijection. The resulting name is globally unique: no star and galaxy, however distant from each other in the sky or the database, will ever share a name.

Stability guarantee

The fixed seed (_SEED = 0xA57E4321) must never be changed. The scatter multiplier (_MIX_A = 316,227,767) is computed at startup from the pool size and is also stable as long as the pool is unchanged. Adding new roots to the pool would shift every name and must be treated as a breaking change requiring a full re-run across all tables.

9. Sample Names

The following are real output from the deployed CORE engine — not hand-crafted examples. Each name is the deterministic result of its object's global rank passing through the bijection.

Comparison to historical star names

Stars — actual engine output

Galaxies — actual engine output

Pulsars

Black holes