It might be more useful to have this discussion in a separate thread, not
tagged on to a discussion about test data sets.
That being said, OLC is defined
here:
https://github.com/google/open-location-code/blob/master/docs/olc_definition.adoc
*Codes with odd length <11:*
Surprisingly, the
Is the code length strictly defined? I would expect that as long as the
string contains a "+" anywhere in the code, then any length (including odd)
should be acceptable (provided the "text location context" such as "Hawaii"
or "Honolulu" makes the code unambiguous). Maybe a degree symbol (°)
The existing test set (encodingTests.csv) already includes codes where one
or both coordinates are negative. There is also a set of tests for validity
(validityTests.csv).
We're not aware of any uncovered edge cases (the main edge cases are at 90
degrees latitude and -180/180 degrees longitude),
