OK this is a research group, how about someone tries an experiment?
I define the problem here to be 'lost emails', not false positives since
some challenge response vendors are already making the dubious claim of zero
false positives on the grounds that the lack of a response means that the
message must have been spam.
In the experiment I propose a number of users install a challenge response
system. We measure the following parameters
1. Number of messages whitelisted
2. Number of messages challenged
3. Time taken to respond to challenge (as histogram)
4. Proportion of legitimate emails where no response was given.
The outgoing message would be the standard one.
After running the experiment for some time the correspondents who did not
reply to the challenge are contacted, are told about the experiment and
asked to fill in a survey on why they did not respond, one of the questions
being how many such responses they have been presented with in the past.
The experiment could then be rerun with a modification to the outgoing
message stating that the message being challenged had been identified as
potential spam and that the message was being chalenged in that particular
instance.
It seems to me that it should be possible to get this type of thing past
human subjects study board, the information being gathered does not relate
to the individuals, does not expose personal data and the study is for the
long term benefits of the subjects.
Phill
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg