andgineer / echo-words / 33672709268

96%

push

github

A wording written in two scripts is counted in both

The usage search matches an exact string, so one script counts a fraction of what a
wording is used in: `прозор` occurs 1676 times in the Serbian encyclopedia against
`prozor` 170, and `сврака` 87 against `svraka` 3. A note carries whichever script the
answer wrote, and answers to Cyrillic submissions do come back in Latin, so without
this a real Serbian word could return the number a coinage returns.

The counts are now added, and the spec says what the number is and is not. It is not
comparable across languages — the same bird is `сврака` 87 against `magpie` 5315 and
`Elster` 6416, which measures the size of three encyclopedias and nothing about the
word — and it is per exact string, so inflections are counted apart from the lemma.
Only nought against not-nought separates anything measured, and even that is shown
to the reader rather than acted on.

A script that fails to answer leaves the count the other one gave: extra evidence is
not a precondition, and losing it must not turn a wording the encyclopedia has into
one it says nothing about.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Rvnc85GEKGv8SdfFhxVQqt

21 of 21 new or added lines in 2 files covered. (100.0%)

2780 of 2931 relevant lines covered (94.85%)

0.95 hits per line