I haven't looked at the code, but I'm guessing it indexes all the Unicode names of characters, and allows lookup by any substrings of those names, right? Maybe using a suffix array. But it would be nice if there were wildcards. Regular expressions would IMO be overkill, but s.t. like Unix globs might be nice, e.g. "a * acute" to find "LATIN CAPITAL LETTER A WITH ACUTE" or "LATIN SMALL LETTER A WITH ACUTE.