On 12/2/20 12:00 AM, Phillip Hallam-Baker wrote:
The biggest mistake in the Web protocols was accepting the dictum of being permissive in what is accepted. Horrors like MIME sniffing might had been avoided if enough browsers had insisted on literal interpretation of content type.
The fundamental problem there is that the server doesn't have a reliable way of knowing the content-type. The server can't reliably tell it using filename suffix. And since the client is blamed when it displays content incorrectly, the pressure is inevitably on the client to fix things. I don't think this could have been avoided by different language in the HTTP specification.
Keith
