Part of me says you should just use machine learning... i.e. Repeatedly train a Bayesian classifier that given an email's entire content decides which authority it is too. And have an interface for requesters to redirect emails that go to the wrong place.
Most of the time it would classify based on the request id in the email address. But then it would learn more fancy things as appropriate, automatically. Hmm, or a decision tree. Anyway, the kind of machine learning is a separate issue, you can feed it into Kaggle... Francis On Wed, Mar 14, 2012 at 10:14:19AM +0000, Robin Houston wrote: > I like that idea. Maybe we should give it a try. Any good reason not to? > Presumably we would discover fairly quickly if it caused more spam-canning > problems. > > Robin > > On 14 Mar 2012, at 09:52, Alex Stapleton <[email protected]> wrote: > > > On 14/03/12 08:26, Mark Goodge wrote: > >> My guess is that what's happening is that they've taken the one-time > >> address generated by WDTK for the first request, and then stored that in > >> their system as your address, and any subsequent requests from you are > >> being replied to that address. Or, more simply, whoever is replying is > >> just looking for "Paul Perrin" in their Outlook address book and taking > >> the first match. > > (I am making an awful lot of assumptions about how things work currently > > here but...) > > > > What about instead of sending the message from > > > > "Paul Perrin <[email protected]>" > > > > You do > > > > "Paul Perrin [Title of FOIA Request] <[email protected]>" > > > > Obviously the responder can still screw up and choose the wrong return > > address, this way does seem like it would make it easier to do the right > > thing using the erroneous work-flow you described them using. > > > > _______________________________________________ > > developers-public mailing list > > [email protected] > > https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public > > > > Unsubscribe: > > https://secure.mysociety.org/admin/lists/mailman/options/developers-public/robin%40mysociety.org > > _______________________________________________ > developers-public mailing list > [email protected] > https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public > > Unsubscribe: > https://secure.mysociety.org/admin/lists/mailman/options/developers-public/francis%40mysociety.org > _______________________________________________ developers-public mailing list [email protected] https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public Unsubscribe: https://secure.mysociety.org/admin/lists/mailman/options/developers-public/archive%40mail-archive.com
