Trackbacks, Queries, and Encoding
This morning, I got a trackback from a Korean weblog. Unfortunately, if you look at how a typical trackback is sent, you see that character encoding information is not provided.
Something to think about the next time you are tempted to think that you can get queries for free. While both HTTP and XML provide mechanisms for defining encoding, support in widely deployed implementations is much better in XML than in straight HTTP.
URI's seem to be converging on UTF-8, albeit at an excruciatingly slow pace. Don't leave this to chance - if you are defining a GenerativeNaming scheme today, make this explicit.
If you are defining a protocol based on HTTP POST, encourage the use of the charset parameter on the Content-Type header. Require it if you can.