Tuesday, June 17, 2003

Airlines use Soundex to match passengers with terrorists

The SF Chronicle reports that airline res systems use an ancient name indexing system that guarantees masses of false positives.

No-fly list ensnares innocent travelers


Many airlines rely on name-searching software derived from "Soundex," a 120- year-old indexing system first used in the 1880 U.S. census. It was designed to help census clerks quickly index and retrieve sound-alike surnames with different spellings -- like "Rogers" and "Rodgers" or "Somers" and "Summers" -- that would be scattered in an alphabetical list.

Soundex gives each name a key using its first letter and dropping the vowels and giving number codes to similar-sounding vowels (like "S" and "C"). The system gives the same code, L350, for "Laden" and all similar-sounding names: Lydon, Lawton, and Leedham."

Soundex is well known these days to genealogical hobbyists. Here is an article laying out the Soundex coding scheme:

"Every soundex code consists of a letter and three numbers, such as W-252. The letter is always the first letter of the surname. "

The Soundex code for my last name is F-624.

Here is an automatic Soundex calculator.

It was a great system when one had to index long lists of names by hand. It was designed to group large sets of names together to reduce workload. It's madness to use it in a modern computer system to check for matches with the names of terrorists.