- From: Larry Masinter <masinter@adobe.com>
- Date: Mon, 4 Mar 2013 14:19:41 -0800
- To: Bjoern Hoehrmann <derhoermi@gmx.net>, Robin Berjon <robin@w3.org>
- CC: "www-tag@w3.org List" <www-tag@w3.org>
I think there are many technical issues but also underlying the MIME content-type vs. magic number sniffing discussion is the political (and thus difficult to resolve) issue of distributed extensibility. The ability to introduce new MIME types and deploy them was an important distributed extensibility mechanism of the web. IT would allow introduction of new types -- proprietary or open or somewhere in between -- as experiments to be tried and, if popular, standardized. I think part of the disconnect is whether this is a good idea. Third party content-type handlers .... are they desirable or not? registerContentHandler doesn't seem to allow you to register magic numbers so if you're relying on magic numbers to determine file types how is RegisterContentHandler to know when the content should be handled?
Received on Monday, 4 March 2013 22:20:21 UTC