This came in 2006 after attending a talk on bioinformatics. I had the idea of making an email client that would take the methods of bioinformatics and apply them to spam-detection.

Searches through input and outputs sequences that are repeated. Because it's intended for text files, control characters are ignored.

FindPatterns [filename] [-b] [-e] [-i] [-o] [-v] [-m<n>] [-l<n>] [-g<n>] [-?|h]

filename

Attempt to read input from this file, otherwise uses stdin.

-b

Keep a buffer to count repeated matches (!o -> b.)

-e

Echo input.

-i

Case-insensitive (not implemented.)

-n

Don't display matches at the end.

-o

Output matches immediately as they are found.

-s

Silent mode - plain output with no extra characters.

-v

Verbose comments while outputting.

-g<n>

Set memory buffer granularity to the closest power of two lower than <n> bytes (default 1024.)

-l<n>

Set match limit to <n> matches (default 4096; 0 -> no limit.)

-m<n>

Set minimum match length to <n> symbols (default 3).

-?|h

Display this help screen and exit.

Adding -<s>- will turn off switch <s>.

Also included is a simple KillSpam email client that takes the patterns generated (from FindPatterns) and eliminates all the emails that have matching patterns.

..

Programming.

Find_Patterns-1_0.zip (27 KB)

Source code for FindPatterns and KillSpam.

killspam.exe (62 KB)

The mail client for Windows console (included with the source code.)

spamregx (1 KB)

Sample regX.

From http://neil.chaosnet.org/code/FindPatterns/.