When I look at web marketing data (e.g., Alexa), I only ever see results for URLs with Latin-based URLs, even for regions with non-Latin alphabets. Do users from places with non-Latin alphabets type in Latin website names, or do they use non-Unicode versions?<p>I know China likes to use numerical URLs, which avoids this problem, but what about other places?
There's a solution but it's awkward. We scandinavians use æøå but rarely do we use them in URIs. Like the electronics chain "Elkjøp" uses "elkjop.no" - this is coincidentally one of the few that has actually registered "elkjøp.no".<p>I have "kråke.re" myself but the DNS entry is really xn--krke-roa.re because international character DNS is an ugly, ugly hack.
i once tried the unicode urls and they are a pain in the a§§ because some browser (and mail clients) interpret them differently. Also some search engines and/or crawlers interpret them different (mostly they are double encoded). This in turn results in several "errors" that the developer then has to re- en/decode again server side to serve the correct content to url.<p>On the other side: Just "transliterate" a url is super simple and people all over the world can at least read the url (and probably memorise). For example: ä => ae => everybody in Germany knows how to read and interpret this.<p>SEO wise: No difference at all.
I believe because of Cyrillic URL spoofing from a about 10-15 years ago most people stay away from UNICODE urls. China's numeric URL's maybe so that tracking URL's are easier, but that's just a guess.