"All the jocks think about is sports, all we think about is sex" -- Revenge of the nerds, paraphrased.

The Purity test is a series of yes/no questions that attempts to gauge your personal purity.

My edited version of the purity test (I just fixed typos and such from the standard version, I didn't write my own).

The following programs analyze a database of purity scores found at this site. The timestamps on the page are not Y2K compliant, BWA-HA-HA-HA. The timestamps in the database are.

The database is a list of about 130,000 scores of everyone who has taken the test at that site. It is a wonderful database, it includes score, timestamp (dayofweek-month-day-hour-min-sec-timezone-year), and the host of the test taker. Unfortunately, the site is sometimes really slow and unreliable so I have cached a local copy (bzip2 (2.1MB)).
Note: the database as is has one really long entry that screws up these programs, it is one of the 37%'s, I think. Simply delete or shorten this entry to fix it.

A program written in C++ (source code for you to laugh at [it works dammit, leave me alone...]) to extract scores. A better way to do this (and all the other programs on this page) would be with some standard UNIX programs like "grep", "cut", "sort" and "wc", which I didn't know at the time.

The results of this extraction from early September 1999. You will have to make the graph yourself, but it is worth it. It's a pretty nice looking curve, off center bell shaped, peaks at about 64% and has a mysterious plateau in the 80%'s.

If you want to know WHEN people take the test, here is a program to extract timestamps from the aforementioned database. (Source code)

A program to smooth out the data from the previous program to a user entered degree so you can get a nice looking graph (source code). Turns out the most people take it at midnight and the least at 6 a.m. Not too surprising, but the shape of the rest of the graph is interesting.

Why have I written these? Because I am me. Why did I put them on my web page? Because anything was better than Russian homework.

A good source for more purity and pseudo-purity tests: The Armory.
My own score.

It has been pointed out that if for some reason you actually want to compile my programs, it would be nice to have the non-standard ap classes so you don't have to rewrite them, so here they are, along with some other classes I have found useful, such as Rando and SysTimer.

Back to home