dmelo / amuzi

Music amazingly free.
amuzi.me
GNU General Public License v3.0
36 stars 18 forks source link

Improve search fuzziness #168

Open virtualritz opened 10 years ago

virtualritz commented 10 years ago

For example:

Enrique Rodr

correctly produces a bunch of suggestions for the late Argentinan musician Enrique Rodríguez. But typing:

Enrique Rodri (one more letter, the ‘i’)

makes the list go empty because Amuzi considers ‘i’ different from ‘í’.

A lot of letters can also be written out, for example, in German ‘ö’ (umlaut) = ‘œ’ (ligature) = ‘oe’ (two letters). Same for ‘ä’ and ‘ü’. ‘ß’ = ‘ss’ etc. Also of note: some people who do not know any better, particularly in the US, use ‘B’ as a replacement for the German sharp ‘ss’, aka ‘ß’. So a guy called ‘Haußman’ (which could also be spelled out ‘Haussman’ or would become ‘HAUSSMANN’, since there is no capital version of ‘ß’) may be errornously spelled ‘HauBman’ by someone from outside Germany who are using a non-German keyboard layout, when searching for them.

When searching for a musician called ‘Müller’ on a non-German keyboard layout, most people would type ‘Mueller’ which is also how this name is spelled on credit cards and the person’s passport.

Same goes for someone searching for a Hispanic musician (e.g. Enrique Rodríguez) on a non-Hispanic keyboard layout. Barring the availability of ‘í’, they'd use ‘i’ instead.

Google is quite good at getting this right btw.