Plain character sorting often puts item10 before item2 because it compares the first differing character. Natural sorting recognizes digit runs during comparison, so the numeric portion ten follows two. ListClean combines numeric comparison with the active interface locale and exposes ascending and descending choices. It never parses the whole line as a number or promises spreadsheet column semantics.
Numeric runs change comparison, not content
For item1, item2 and item10, natural ascending order matches how people usually read numbered labels. The original strings remain byte-for-character unchanged apart from any separately selected trimming. Values such as 02 and 2 may compare as equivalent numeric keys; ListClean then uses original position as a tie-breaker so repeated runs remain deterministic.
Alphabetic order follows the active locale
Accented and non-Latin letters do not have one universal order. The browser’s reviewed `Intl.Collator` data applies the active EN, RU, ES or FR locale, with accent differences preserved. Changing the interface locale may therefore change a multilingual list’s order even when the input is identical. Use the locale that matches the list and inspect the final sequence.
Reverse is not descending sort
Descending order compares every pair and produces the opposite natural ordering. Reverse mode simply flips the surviving current sequence. If the list begins B, A, C, reverse produces C, A, B while descending natural sort produces C, B, A. Keeping both operations separate avoids pretending that a positional reversal is an alphabetic result.