From: Russ Allbery (rra@stanford.edu)
Date: Mon Feb 02 2004 - 23:27:31 CST
J B Moreno <planb@newsreaders.com> writes:
> Depends upon how you look at it. If you think it just means "contains 8
> bit data" or "an compiled program" then you're right, but if you think
> of it as "a file included in a mail/news message which may or may not
> have 8 bit data" then it was.
When you use that wildly unusual of a definition of a common term, it
would save a lot of time and confusion if you would mention your unusual
usage.
> Regardless, however you look at it, the Subject is only mostly
> unstructured.
That's like being kind of pregnant. Either it's unstructured or it's not.
We've seen the complexity that results from trying to make it structured,
not to mention the fact that e-mail doesn't make it structured, and the
only thing it actually appears to accomplish is to quiet the fears of
people who think that there's no other way to get rid of Sv: beyond making
standards unnecessarily complex.
Not declaring Re: to be a protocol requirement isn't going to mean that
software is going to magically stop generating it. That's not how people
use standards or write software. It clearly has to be documented
somewhere, but it doesn't need to be in the ABNF and it doesn't need to be
a protocol requirement. The protocol does not break when people omit the
Re: in followups in software that does include References.
Declaring the header unstructured doesn't mean one can't put information
in it, like the binary groups do. It just means that the standard isn't
helping them find the right structure and that if a lot of people want to
use the same structure, they should probably find a more machine-parsable
way of communicating that data. Like References.
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>