New project to search for missing people

[rainbowgirl28]

A friend of mine is working on a new project that he hopes will help find missing people. Check it out http://forge.novell.com/modules/xfmod/project/?pulpfile

I am a programmer/network administrator with a penchant for data
mining. I was taking a vacation with my wife and kids this past week
at a local Delaware beach (Bethany). As such, I bought a copy of
Wired magazine to read on the sand. I read the article in the August
issue entitled "Raising the Dead". This article was the final piece
in a puzzle I've been trying to solve for a few years now; but more
about that later. I was captivated by the idea that you guys were not
only using the internet to find missing persons or solve cold cases,
but also that you were seeing some success!

Here's where my puzzle gets solved: what if I have an idea that would
significantly spead up your search time by allowing you to use your
computer or some other computer to do most of the searching for you?
I'll explain my puzzle and you may see why this could be something good.

Have you ever heard of popfile? It's an open-source project based out
of http://popfile.sourceforge.net that blocks/filters spam.
Basically, you tell it what messages are spam and it does statistical
analysis on each method based upon the words in the message to
determine whether to classify it as spam or non-spam. I've been an
avid popfile user for years and LOVE it. It's 99% accurate inside of
a week of use! And it stays accurate. I was so intrigued by this
software that I started thinking up other uses for it. One of the
uses I came up with was a sort of auto-browser that surfs the web for
you and compiles a list of links you should visit when you get home
from work. The idea being that you tell it what you like to read, run
the program, and it creates a digest of what's on the web today that
you may like. It creates a pulp of the web for you, if you will.
The problem was that I couldn't get much interest for such an
application. There didn't seem to be much of a use for it in the
mainstream or even niche markets. So my puzzle became this: you have
a viable idea but where does it fit? Where will it scratch the
biggest itch?

Clearly no itch could be bigger than that of unavenged murders and
lost loved ones. Pulpfile (as I've come to call it) can scratch the
missing-persons/cold-cases itch as follows:

1. User trains it to look for a specific person based upon whatever
information is on hand at the time. As more info is gathered (by
Pulpfile) and corroborated (by a human), it is added to the training
profile so that Pulpfile gets "smarter".

2. Pulpfile searches google.com, groups.google.com,
images.google.com, along with missing-persons/cold-cases sites and
law-enforcement databases (all online).

3. Any pages/graphics found that seem to be a match get put into a
file that is a list of links for the human to go through.

4. Multiple missing-persons/cases can be searched for at one time.
Pulpfile will keep them all seperate and mark each link accordingly.

Clearly this won't remove the need for human interaction in solving
cold cases. But it can augment the search and, at best, speed
searches to a fraction of their current standard.

If you are a perl programmer or even just a C/C++ programmer willing
to learn perl, please join this project so we can give the cyber
detectives a helping hand and ease the suffering of loved ones that
much quicker: http://forge.novell.com/modules/xfmod/project/?pulpfile

This project is open-source and will be released under the GPL or an
equivalent license. I never do this kind of work for money.
--

Regards,

Bryan Simmons