As you know I have my little test site on Appharbor http://vidyano.apphb.com/#!/KanjiOverview

The Kanji data came from a SQL server database with an order by on the Kanji character. But once I cached the query (automatically refreshed every 30 minutes) the order was completely different. It seems that ordering and unicode don’t really match, apparently Japanese doesn’t really have a fixed (or at least a single) way of ordering their characters. And now because the order by was happening in .NET it gave me different results.

The only thing I could find was using special comparers, all linked to the ja-JP culture. But then it hit me, my current thread’s culture was already set to ja-JP because I was playing with the Japanese date format. Changing the thread back to en-US gave my back the same results as before.

The order is now uses is by radical and then stroke count which makes it easy to see the related/same kanji.

Does anyone know what order the Japanese culture info uses?