Archive for 25 November 2005
Why Don’t Nordic Characters in URLs Work in IE and Opera, But Only in Firefox?
25 November 2005 in Root | Comments (0)
I noticed a strange thing the other day: I tried to download a file with Nordic (aka. Scandinavian) characters in the name and met an error page. However, when someone else used Mozilla Firefox to download the exact same file, they were able to download it! 
This prompted me to believe that it had something to do with Opera’s IDN whilelist and I henceforth modified it to allow Finnish domains (www.***.fi). But, alas, the result was the same: error 404 [see pic above]. 
Well at home, I tried the same experiement using Firefox, Opera and Microsoft IE but this time while running the requests through Fiddler HTTP debugger proxy. I tested the requests to a file named: åäö.html and this is what I found out:
When I try to open åäö.html in Opera and IE, the browsers encodes filename as %C3%A5%C3%A4%C3%B6.html, which the Linux machine (that my website’s running on) doesn’t identify as an existing file and thus returns the 404 page. On the other hand, in Firefox, åäö.html is encoded as %E5%E4%F6.html which the Linux machine properly identifies as an existing file and returns the correct page.
According to W3schools, the proper URL encoding for å, ä and ö is the one that Firefox produced, whereas the result from Opera and IE is completely off the chart. It is a bit sad to see the Norwegian company behind Opera not noticing this (or have they?), but on Microsoft’s behalf, I could have been expecting it (they are an American corporation nonetheless, which probably doesn’t use Nordic characters much).
I guess this is one point that shows how open source can be a bit ahead of the game (read: corporations). 
UPDATE: See also the continuation “Nordic Characters in URLs — Part 2“