h a l f b a k e r y
i v n i n seeks n e t o
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
This is an email filter that catches multiply-forwarded messages and checks to see if you have received them already. If you have, and if the email contains no additional content aside from the new headers, it deletes the message.
I frequently receive many copies the same email message forwarded
via different lineages of co-workers. Deleting them by hand is too old fashioned.
||Idea: quantum linked emails that can be rated by all recipients. If an email has received less than a 2 star rating from previous recipients it will be blocked. Constantly updates itself (via quantum linkage) to account for the ratings of people who received the email after you did.
||Sometimes the value added to the information is precisely from whom it is forwarded. Coming from the administrative assistant to the CEO is a dfferent story than coming from your cubemate.
||Sorry, this idea got filtered out ...
||What's wanted, here, is software that
||1) recognizes overlap (shared common text) between
emails, and groups them.
That part's not hard: easy
to generate a correlation matrix for everything in your
inbox, then apply a statistical clustering algorithm (there
are lots; determine empirically which works best). Then,
in the listing of emails received, replace all the emails in a
cluster with a single entry.
||2) Display that entry as a graph (in the mathematical
sense: a bunch of nodes, joined by edges): there's lots of
software, these days, for displaying that sort of data.
Each node would be labeled with the sender. Each edge
would correspond to a congtiguous block of text common
to two nodes
What to do when the same block of text was common to
more than two nodes? Represent that text with its own
node (different graphical representation from emailNodes
-- maybe a smaller differently colored dot) and have edges
going from that textNode all the emailNodes that have it
||3) After that, the sky's the limit. As a practical matter,
you probably want to be able to step quickly through each
email in the cluster. But think of all the other cool stuff
you could do. For example, hover the cursor over a bit of
text in the email, and have the graph-display rearrange
itself so all the nodes that share that bit of text are
grouped together on the screen. Arrange the nodes
chronologically, with the earlier ones closer to the top of
the screen, so you could see how information flowed
through the organizational hierarchy.
||In fact, I'd bet Microsoft, Google, and quite a few others
are developing, or have developed tools like this, for
offline analysis of email corpora by corporate managers, IT
folks, sociologists, and law enforcement agencies. But I'd
like to see them used in real time, on a small scale, by
ordinary email recipients.
||... sorry, got carried away, there.
||Then again many mail programs cannot even manage not to put
"FWD: FWD: FWD:" in the subject line of a thrice forwarded message.
||I wouldn't hold your breath for an industry standard or even de-facto
standard anytime soon, but if you choose an opensource mail reader,
you might be able to talk developers into adding the feature. Or add it