This weekend's project: shellcompanyfinder.com

25 points by fnid2 ↗ HN
The thread that started it all: http://news.ycombinator.com/item?id=1134813

I am soliciting ideas for shellcompanyfinder.com I'm going to build it this weekend. I'd love to have some help with it. It's a great idea and a way that we can start to combat some of the ridiculous behavior from corporations who hide their actions with shell companies.

For now, it's a list of companies with a many-to-many to associate them. We can add comments to companies to discuss whether or not they are actually shells. It also accommodates actions these companies have taken so we can roll up all the bad behavior for the uber parent "name brand company." These types of tricks aren't just perpetrated by corporations no one has ever heard of. Think Enron, MCI/Worldcom, ATT, and a host of others...

You can also vote on whether or not you think a particular company is a shell and provide evidence in the form of uploaded attachments that could be scanned documents or public notices for DBA's, etc.

Please share your thoughts. I want to have something up by Monday morning, so the sooner the better!

EDIT: The site is live now: http://www.shellcompanyfinder.com though obviously still a work in progress. If you are interested, create an account. We may want to set up a gotomeeting to run through how to make changes. Everything is behind the login wall right now.

11 comments

[ 3.7 ms ] story [ 41.1 ms ] thread
I've begun a data model at yuml, you can see it here:

http://yuml.me/diagram/dir:lr/class/%5BActions|id%3bShellCom...

The schema looks good. I'm a web developer and would like to help, what do you need?
Awesome! I'll get the website up soon and post a link. We can modify it over the web with html/css/js/sql with the web platform I'm going to use. keep checking http://www.shellcompanyfinder.com Should be up ... hopefully soon.
OK, I see it. If you'd like to communicate over email you can reach me at noblethrasher@gmail.com
I created an account for you. Should see the email soon.
What's the point of the createid fields? And why do you hardcode two related companies instead of going for a N-N relationship?
The createid is the unique id of the user who created the record.

There is an N-N relationship between companies persisted in the RelatedCompanies table. Perhaps I should add a field to the RelatedCompanies table that is a description of the relationship: Subsidiary, Partnership, Spin-off.

When we start, we are probably going to start listing companies without really knowing how they are related. They are just a group of companies. I considered a self-join, but we may not know who the parent is and a company may be related to more than one company through some mechanisms I don't understand yet.

you will want to take a look at how wikileaks does it thing. you will need people in asset havens like bahamas/etc uploading you stuff on the sly, as my understanding is a lot of the way this works is simply information hiding by having an official registry that is text only, having legal standing, not searchable in any electronic way. when a crime is committed you get to initiate a legal proceeding that will eventually unravel the paper trail in the bahamas or wherever, but each step takes n amount of time, and when you have shell companies under shell companies the stalling strategy can be quite effective.
It's going to be quite difficult. I am also concerned about the legal issues I may face. I don't doubt we'll get C&D's and other legal threats for defamation or libel.

I really don't know what I'm getting myself into here, but I feel drawn to it.