The HTTP Server doesn't always give the right mime information back to the
client.
There has been a few attempts at simple methods to determine the right
mime to use though none have been really right.
The goal is to write a new class for determining the mime output.
It will try using a few methods to get the right result.
First it could use the registry enteries to try and find a match, if that
fails then it should have its own list of known mime types and compare
against that.
Finally if those two methods fail it could use the other method already
implemented that reads the first 256 bytes of a file to try and determine
the mime.
Anyway more research needs to be done.