Re: Signed message and public open injectors [was a completely indequate title]

New Message Reply About this list Date view Thread view Subject view Author view

From: Brad Templeton (brad@templetons.com)
Date: Tue Jan 21 2003 - 14:50:30 CST


Signing is not in the spec, but the question of header ordering is.

It is not yet decided how we will sign, but it leaves more options
open to the designers of signing systems if the current spec says
that news tools must not re-order headers, and further that if
they add headers, they should put them at the top.

Now as it turns out the news transports today do not re-order headers
so they have nothing to do to comply. They do insert Xref
headers (because they are twinned with the local news database)
and it would a nice idea if they slowly started moving this to the
top as well, but clearly for now we just special case the Xref
header when it leaks out of a site.

However, over time, injectors might get slowly redeemed. Fortunately
the injector <-> user relationship is one to one. People can base how
they compose their articles based on the rules (and bugs) of the
injector they plan on using.

So while avoiding debating signing (though it is a refershing break
from the wars over UTF-8) the question over whether this draft
should require future news software to change the header as little
as possible and in nice ways is still apropos.

Of course, injectors that want to rewrite headers (rather than just
add them) are irredeemable when it comes to user-signed articlers.
Fortunately that's not an issue as we will just bypass these
injectors.


New Message Reply About this list Date view Thread view Subject view Author view


This archive was generated by hypermail 2b29.