4 ms·
This Is America, Take Your Unicode Somewhere Else
- noss 17y agoThe experience and the conclusion do not match. If seaching for "bjorn" doesnt match "björn" his little search engine just needed to know to turn "ö" into "o" and add that as a keyword as well for the record. Some users want "torbjoern" added as well, as that is how they transliterate "ö" when the keyboard lack it. Unicode actually include lookup tables for locale dependent collation rules. English or American people expect "Östermalm" to show up among words beginning with O, but we Swedes do not.
- allenbrunson 17y agowherever there's a silver lining, you can be sure ted dzuiba will be along shortly, to tell you about the cloud that goes along with it. why would anyone want to live this way, always thinking the worst of everyone around him?
- dan_the_welder 17y agoI thought he was funny, in my favorite cynical way.
- ggchappell 17y agoPerhaps the real problem here is that, conceptually, o-umlaut (say) is not really a separate character from "o". Rather, it is an "o" plus an accent. Typographically, this is obvious; how to store o-umlaut in a text data structure is not so obvious. Unicode takes the view that it is a separate character, and perhaps this is the wrong view. Putting it a different way: the problem isn't that diacritic marks are bad, nor is it a clash of cultures. The problem is that the Unicode standard does not match up well with the way people actually use character strings.