--- In Anti-spam@yahoogroups.com, "Capri" <capri@...> wrote:
> "- Next require the new ID pass some sort of Turing
> test, not a Captcha that can be cracked in a fraction
> of a second, but something that demonstrates a real
> human comprehended a question and thought about an
> answer, maybe did a Google search to find the answer,
> and entered the correct answer, which takes a minute
> or two for most people to complete."
> --
>
> Amen! That would be thankfully a user-friendly way of blocking spammers
> without punishing users with visual impairments the way that awful captcha
> code does.

Ah, glad you like my opinion (anti GIF captcha, pro something else).
Have you seen the many samples posted to various newsgroups?
Search Google Groups for the string:
"Nobody in their right mind likes spammers, nor their automated assistants."
which is a preface to the various alternatives to GIF captcha.

> --
> "It will take some programming to implement such a system.
> Is there anybody interested in working with me on design of such a system?"
> I would if I was any good at programming, but would be glad to test it out
> when it's been realized.

You don't have to be an expert at programming to help *design* the overall
system. For example, which of the many different types of short-answer
Turing tests do you like best? What sequence of tests would be optimal?
Under what conditions should the system suspect an automated new-account
attempt and consequently impose stricter more difficult tests?

> > nuala-alexander
>
> Yeah, I got some of those, but I got a bunch of others
> too. As soon as I banned a lot of nuala-alexanders, I
> got a shitload of some completely different name
> spamming in her place. Most of the spam was for Windows
> Vista piracy, and one other scam I forgot. Same with you?
> --
> There are a ton of those addresses, junelle-coyne savanna-terk and
> sherry-heath have started up again after a long silence. Now they have new
> addresses to ban.

Yeah, I've gotten a lot of those in my Yahoo! Groups too.

By the way, something that would be helpful for those, which I'm considering
setting up: (1) daemons that watch for specific kinds of events, such as
spammers posting to a Group. (2) alert systems to people who might be
interested in prompt notice of such an event, such as owners/moderators
of the groups. (3) cross-links from 1 to 2, so that when a spammer hits
a particular Group all the owners/moderators who have personal alert systems
can be notified. Are you interested?

So how would the daemon know that a message to a Group is spam?
First, if it's from one of these already known patterns of spammers,
such as you listed above.
Second, we could have a convention that if anyone in the Group posts
a Reply that just says SPAM in the body, that would be a trigger for
the moderator to be notified of whatever message it was in reply to.
Correct SPAM markings would quickly ban the spammer.
False alarms would get the false alarmer banned instead.
In any case, the owner/moderator wouldn't have to check the Group
constantly, which is a royal pain. Instead the daemon would catch
repeat-pattern offenders, and the first regular member of the Group
who sees the spam would catch new-pattern spammers or one-time spammers.
So the owner/moderator wouldn't have to check the Group hardly ever,
except when the daemon or a regular member spots a new spam.

So how would the daemon see the messages in a Group? The daemon would
have several Yahoo accounts of its own, log in automatically and go to the
groups to see what new messages have been posted. Setting up new accounts
and joining groups would be done manually just once per daemon-account
or per group respectively. I might automate some of that, so that
anyone owning/moderating a group who wants to use this service would go
to a Web form to verify his/her e-mail address and say which Groups
need watching, and then I'd get an alert saying that somebody needed one
of my daemon accounts to watch that group, so then I'd manually do the
necessary stuff.