Re: Proposed EU Directive on Electronic Commerce

[Dave Wilson <]

>> Labelling anything at all is a terrible idea in the first place, it's a
>> non-stop ride to the pits of despair :-)
[snip]
> Wait a minute... censorship?  The idea was that the *spammers*
> (advertisers) would label their messages so as to enable ISPs to pass them
> through to customers that have indicated they did want to receive UCE 
> marked with those specific keywords
[snip]
> Why is this such a bad idea?  Give me an example of how it can
> fail miserably because I can't think it up myself.

Labelling carries with it an implicit narrowing of the audience. I can't see
spammers swallowing this voluntarily; their forged return addresses suggest
that professional "we-spam-4-u" crowd don't even care about keeping their
address lists up to date. I see HTML <meta> tags, and their subsequent misuse,
as an established equivalent.

However, I may be wrong, and I'll concede that there's no harm in making room
in the specs for a ***strictly voluntary*** categorisation system.

A legal requirement to label spam suffers from the same problems as any global
labelling system. The primary problem is that global tastes are way too
diverse for a global system. What causes outcries in the U.S. doesn't bat
eyelids in Japan, and vice versa => no middle ground. What global body would
administer the categories? What court would have final say on a particular
categorisation?

This can't even be split among countries; once separate authorities are
involved, it becomes too difficult for either the end-user or the spammer to
work with, and the consequences of the spammer choosing not to work with the
system may not be as benign as we would like to think.

There's more. In order to reduce the biggest problem - threat to our servers -
the accept/reject must be pushed as close to the source as possible. At the
client end is useless. But added complexity in the decision process,
user-configurable I might add, will not sit well in our MTAs.

I'm making the assumption that our primary problem is threat to our network
infrastructure; a simple yes/no at the ISP's MTA is **probably** just about
close enough to the source to fix this (but make no mistake, it's still a
hack. Though I'd prefer no spam at all myself, I just don't think it's a
realistic target).

If one really wishes to receive updates on a particular subject, there are 
many, many better directories and opt-in lists that are much more reliable,
with a much finer granularity, than any massmail categorisation system. This
isn't a cop out. Mass mail has its place. Even targeted mass mail is a
sensible concept. But it is not up to us, and our busy servers, to do that
targeting.

Negotiating the preferences of the user is a duty that must directly involve
the sender of the message. This is a duty that mass mail of any kind has never
really chosen to accept; email more so than any other. It would be a mistake
for ISPs to offer to accept this responsibility instead - and even if we did, I
doubt the offer would be accepted in any other than the most cynical manner.

Regards,
Dave

-- 
----------------------------------- dave.wilson@localhost  Dave Wilson, HEA-NOC
HEAnet Limited, Marine House, Clanwilliam Court, Dublin 2  ph.  +353-1-662 3412