Re: taxonomies2004-02-07 11:47:25
This is a general comment without specific reference to any individual
comment found so far in this "Re: taxonomies" thread.
In discussing all the ways that a sender might construct a bundle of
information content for a recipient, the first key must be to bundle it
and the second must be to label it. Think of it as Tagging and Bagging!
This is the problem that MIME solved with an open ended "Tagging and Bagging"
Protocol which identified and protected the content, but did not dictate what
the recipient MUST do with it. Recipient freedom of action is critical.
Receiving something that is restricted to only doing what the sender chose,
in some critical sense, violates the entire concept of mail.
When you mail anything to me, it becomes mine when you send it, and it is for
me to deal with the content as I may wish. It is not yours anymore, and your
only control over it is limited to prior "contract" agreements as to what I
will do with it. So, in MAIL-NG we must not violate this rule of ownership.
If I want to use a MAC, and you want me to use WINDOWS on your message, it
will go in the trash and that will be that.
The big MIME fight at the time, was between "JUST SEND 8 BITS" in all mail, and
let the recipients figure out what the content was by examination, vs "TAGGING
and encoding binary into ASCII body parts" with an open system of defining new
kinds of content, with a registry of types that would identify how the content
is to be decoded and rendered for the recipient.
So, you all might be dreaming of discarding MIME, but you cannot do that without replacing it with a new Tagging and Bagging protocol.
Tagging identifies content types by reference to a document that specifies
how it ws packed and how to unpack the content.
So, my main point here is that MAIL-NG should retain the ability to use open
bagging and tagging protocol standards as a primary architecture.