D649 - A name that cannot be confirmed

Status: published Date: 2026-09-09

The first unnameable hash that got an answer

libSceAgc::0x53bbd82b51d172db is imported by PPSA02664, PPSA03416 and PPSA28061, and 30,184 names plus 3.9 billion generated candidates never explained it. SELFish answered REQ-20260909T1250Z-1f74: five independent published identifier tables attribute it to sceAgcInit, and

Nid::of("sceAgcInit") = 0x916dc62dbed07cf8   ≠   0x53bbd82b51d172db

The same tables list both identifiers against that one name. The name does not hash to the value a title actually imports, so no name search can ever reach it. That is why the search failed, and the failure was not a gap in the vocabulary.

It is systematic rather than a typo: of 116 published libSceAgc identifiers, 108 reproduce from their names and 8 do not, five of those being a second identifier on a name whose first one reproduces. Decorations, the empty suffix, and both byte orders against 837,352 words were all tried and ruled out before the answer came back.

Why it does not go in the symbol database

Every StaticSource variant in orbistoun-nid ends the same way - and then the hash confirms it, as every name here must be confirmed. That is the property the whole database rests on: nothing in it can be wrong in a way that produces a false name, because a wrong name simply does not hash. sceAgcInit against this identifier cannot satisfy that, and adding it would replace a structural guarantee with a policy.

crates/orbistoun-names/data/vendor.toml is no better a home for the same reason from the other direction: it is a vocabulary for brute force, and a word that does not hash to the target contributes nothing to a search.

So it is recorded as knowledge, at known_by: published

The knowledge files already have the vocabulary for exactly this: a fact whose provenance is a document rather than a derivation or a console. The entry carries the attribution, its five sources, the hash that does reproduce, and the explicit statement that this project cannot confirm it the way it confirms every other name.

The ceiling is stated in the entry rather than left to the reader. published is the second strongest grade this project has, and it would be easy to read an entry at that grade as settled. This one is settled about what the tables say and about nothing else.

What it is worth

Three titles stall inside this library's initialisation, which corroborates the attribution semantically without confirming it - and that corroboration is itself only suggestive, since a first call is where anything stalls. What the entry actually buys is that a reader meeting the bare hash now knows what it is believed to be, and knows not to extend a word list to find it.

The systematic finding is worth more than the single name: libSceAgc carries a class of alias identifiers, so the next unnameable hash in this library should be checked against that pattern before anyone spends another 3.9 billion candidates on it.