XML/ZIP/OctetStream File Mime Type Disambiguation

[deeplow]

Many file are in fact zip files with a particular structure which is then interpreted by the file viewer. This fact makes these filetypes particularly difficult to identify mime type because they can be multiple file formats. Examples of files that have application/zip mimetype:

• office files
• Comic book zip (.cbz)
• .xps

Then we also have files which are XML-based and are thus identified as text/xml:

• .fb2 And some are application/octet-stream:
• .mobi

My suggestion would be to somehow expose the file extension to the conversion process to help identify the file type.