A user uploaded an image file which had the species name as filename, but he got the genus typed wrong by missing an “i”. As a result, a totally different species appeared on the observation: a plant instead of a dragon fly.
I could reproduce that using the web site with Firefox on Windows:
if I name the file “Bracydiplax farinosa.jpg” instead of “Brachydiplax farinosa.jpg” (in this case, missing an “h”, did not test with the “i”), the result is “Paederia farinosa”. I.e. the “farinosa” part was right, but the genus totally wrong.
See also the discussion on the observation
https://www.inaturalist.org/observations/132143548
Could there be some hints to the user that the match was not perfect?
Or could the algorithm be optimized to look for the best of potential matches (e.g. least number of differences in the name)?