[Dataloss] An amazing use of DLDOS

[Chris Walsh]

Nice.  

These records need a unique identifier to facilitate linkage of information
from other tables.

For example, I have:

address
stock symbol
exchange
NAIC industry code
Date of actual breach
Date of breach discovery
Links to primary sources (NY state reporting forms, notice letters)

for many of these.

Perhaps you can backfill a unique identifier into the CSV file for now,
and when future records are added, they can look like this:

CWALSH-MMDDYYYY-nnn

This way, you will not have any collisions 
(unless another C. Walsh comes along), and you will not need to pre-assign
blocks of numbers to anyone who wishes to report.  Should John Smith and
Jim Smith both decide to get into the act, then perhaps the dreaded "jsmith02"
solution can be adopted.

If someone objects to a tag like 'cwalsh' going into the db, then
they would need to say so.  Presumably, a privacy-conscious group like
this will be able to work through the issue.

This is all off the top of my head as far as the implementation, but I have
thought at some length about the need for an identifier.

Thoughts?

Chris

P.S.  I love how these guys write something spiffy in 3 days. I am eager
to see what can be done with an "expanded" DB.  I "know", for example, that
Google Maps could be used to great effect with this information.  If I could
code my way out of a wet paper bag, I'd be on the case.


On Tue, Sep 05, 2006 at 08:19:40PM -0400, lyger wrote:
> 
> Our friends at mailerblog.com have applied attrition.org's Data Loss 
> Database - Open Source in quite a cool way:
> 
> http://www.mailerblog.com/dataloss/dataloss.php
> 
> If anyone else has any ideas, the raw data can be found here:
> 
> http://attrition.org/dataloss/dataloss.csv
> _______________________________________________
> Dataloss Mailing List (dataloss at attrition.org)
> http://attrition.org/dataloss
> Tracking more than 143 million compromised records in 337 incidents over 6 years.
>