Actually, we're not trying to reinvent the wheel here,
rather adapt it to suit a slightly different purpose.

Below are a few main points and reasons as to why we've created filetype:

    * file does not work so well for loosely defined filetypes ( ie, vCards )
    * file uses a text-based type database which can impose unwanted delays
       in frequently invoked processes
    * file does not have a heirachial type tree (ie, executable->MSDOS->EXE )
    * file is not designed to be incorporated at a source level into existing
      projects
    * Simpler and broader type detection engine ( 'file' is very good at
       pulling out every detail about a file, ie, the resolution of an image,
       however we do not wish to seek out such fine details )