2015-05-04 20:51 GMT+02:00 Richard Wordingham < [email protected]>:
> On Mon, 04 May 2015 10:22:16 -0700 > "Asmus Freytag (t)" <[email protected]> wrote: > > But people writing character pickers really should mine these. > > I agree. However, some don't even give the character name, which > lack can be really annoying with some diacritics. My work around is to > look up the codepoint from UnicodeData.txt. > Ideally a perfect character picker displaying names should allow users to personalize these names, and possibly even save them online to a cloud with user preferences, possibly with a sharing option allowing to feed a database per locale, with votes/ratings, so that this database will progressively be able to return names that have the best agreements. Such tool should of course use a locally cache when it will query names from the shared database, and should also contain an option to update it entirely from a snapshot (just like we perform regular software updates). Such systems do exist for various applications in other domain, e.g. for rating web sites or their security/risk per domain name, or for mail blacklists. It could be used as well for a localized database of character names. The database would also be able to list known aliases (by sorting them in rating order and extracting the top 10). With that system it would be even easier to perform plain-text searches of character names using more user-friendly descriptions, capable of finding related characters, or characters suitable for some usage: the character picker would then list all these characters found by name, or part of their name.
