Technorati vs. Knuff
I was just fiddling a bit more with my MoreSimpleTags WordPress plugin that supports several services (Technorati and Knuff, in my case). I noted in one of my previous posts that the letter ä wasn’t coming out right.
It turns out Technorati and Knuff expect non 7-bit characters encoded differently: Knuff wants them in ISO-8859-1 URI encoding like this vinbär [http://bloggar.se/om/vinb%E4r] while Technorati prefers them URI encoded UTF8, like vinbär [http://technorati.com/tag/vinb%C3%A4r].
D'oh.
Update
Seems I was a bit hasty or not paying attention. Both work with UTF-8. So the code is now urlencode(utf8_encode(html_entity_decode(trim($tag)))).
Update #2
Ok, I snug a utf8_decode in there for good measure too. Turns out part of the problem is that Knuff won’t eat accented characters (which is a bit weird because they are quite common in last names in Sweden).