I spent most of my weekend working with folks from CivicSpace
, and the Social Software Foundation
to develop the People Finder Interchange Format
, which is intended to help the various missing-person databases exchange and pool their information. The Katrina People Finder Project will be using it as the interchange format for all their data.
With the database that Miriam Walker and i hastily put online for September 11, the two biggest problems were the inability to ascertain the reliability of records and the inability to update or remove records. The reliability problem became more severe as i started importing information from other databases; at its busiest, the September 11 database was slurping records from around 10 or 12 other databases, mostly by scraping their web pages.
PFIF provides a standard way to transfer records between databases, while addressing these two big problems. To help establish the reliability of a record, PFIF includes fields to identify the person who entered the data and the source where the data originated. To handle updates, PFIF provides a second record type, a timestamped update record, for posting information that changes over time. PFIF assumes a post-only universe: you can't go back and change something that was previously posted; you can only post an additional update. In a way, PFIF's person-with-notes model parallels a blog's entry-with-comments model.
* * *
The Katrina People Finder Project could use your help.