non-censorous spam control (was Re: Spam is Information?)

[Adam Back <aba@dcs.ex.ac.uk>]

A good post.  Your notice-board analogy was reasonable.

How about this as a stop gap measure: rather than sending cancels for
spams posted to multiple newsgroups, simply modify the distribution.
That is cancel all but one groups worth of articles, and then modify
the Newsgroups line to cross-post to all the groups the article was
originally posted to.

That way you are not as such "cancelling" anything as a net result,
but are just "compressing" your news spool or downstream feed, no
information is lost (other than the information that the poster is
clueless and doesn't know how to cross-post, which is now disguised by
a compression agent trawling through the news spool fixing things up).

Sound reasonable?

If news admins, or others getting involved with issuing forged cancels
are not willing to do this it suggests that they are making judgements
about the content of the posts as well as claiming to want to save
bandwidth.


Another longer term way to improve the situation is to charge some
small token amount per article, just to encourage people to use it
with some intelligence (use cross posts rather than separately
reposting to each group).

It is also entirely possible for people to have 'bots which auto-post
in response to articles matching keywords, or matching authors.
(We've seen a few of these on cypherpunks).

If people want to make a nuisance for others by spewing random garbage
via bots to newsgroups, they could post mega bytes of stuff per day
and swamp the content.  What can you do about this?  Charging a small
amount per post, or per megabyte would provide a small disincentive
for this type of behaviour.  However it would never reflect the true
cost to USENET bandwidth as a whole.

One interesting idea which has been floated on this list in the past
is for authors to have their free posting rate moderated by other
peoples ratings of their posts.

One way to implement this is for other people to pay the author for
their articles a penny if they like the article.  That way people who
write things which others find interesting to read get subsidized
posting.  Is it still free speech if you have to pay for your posts if
you're arguing for an unpopular minority?

Also, this might be an interesting information market model because
technical experts might even find themselves with a well paid job of
answering technical questions in newsgroups.

Adam
-- 
Have *you* exported RSA today? --> http://www.dcs.ex.ac.uk/~aba/rsa/

print pack"C*",split/\D+/,`echo "16iII*o\U@{$/=$z;[(pop,pop,unpack"H*",<>
)]}\EsMsKsN0[lN*1lK[d2%Sa2/d0<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<J]dsJxp"|dc`