-=> Quoting Chris Gunn to Leonard Erickson <=-

 LE> If I was writing it, I'd incorporate it into the same routines that
 LE> handle dupe checking (another thing that gateway programs don't
 LE> handle).

 CG> Since the cancel message will consistantly arrive after the message to
 CG> be deleted has already been tossed, there's not much point adding the
 CG> code. 

 Bad assumption. It is quite possible for cancel messages to arrive
*before* the message. And as an "end node", I want the ability anyway,
simply because the majority of cancel messages these days are cancels
of "spam" postings. 

 CG> The Fido style of messaging is based on a different kind of moderator.
 CG> It's freedom of speach instead of censorship oriented.  If somebody
 CG> doesn't like what I have to say as a moderator, everyone gets to see
 CG> their complaint and I have to use a bit more diplomacy.  If things get
 CG> totally out of hand, I have the option of having the problem poster's
 CG> access to their echo cut. 

You *really* don't understand Usenet, do you?

Moderators on usenet don't issue cancels because all postings in the
(few) moderated groups are *mailed* to them. They then approve the ones
they wish to have appear in the group, and put them into it. This is
*not* "censorship", it's editorial control! The very basis of freedom
of the press.

Cancel messages started out as a way for someone to "take back" a
poorly thought out message. And due to the "flood fill" method of
transport used by the Internet, they usually worked quite well (cancel
messages tend to have higher priority than regular messages for one
thing, and they get carried by most, if not all, sites).

There have been some problems in recent years with forged cancel
messages. However, this is a minor problem. Being able to accept cancel
messages from "trusted sources" to get rid of all the commercial spam
out there is a *very* useful feature.