I found out about this exploit by means of this article in the WIRED Magazine. The “more intellectual” article can be found in the Proc. Natl. Acad. Sci.

If a man has no arms he can’t play guitar.
This man plays the guitar.
Therefore, this man has arms.

Unfortunately, statistics is all about likelihoods.

Today we’ve been looking for videos for class. Next to Bruce Tognazzini’s Starfire this is the nicest video prototype I know - done by Apple back in 1987.

Now while Monkey Island deserves the price for the funniest piracy protection, SPSS (Statistical Package for the Social Sciences) clearly rates on the other end of the scale - it’s a yearly nightmare that haunts you.

SPSS has a proprietary licence and knows a variety of registration codes to validate your product. The cheaper ones allow you to run SPSS for a year, the more expensive ones ($619.00) for a lifetime, and the cheapest (”Student edition”) ones simply disable a couple functions that are essential for research, so you have to buy the whole thing anyway as soon as you get into the second semester of statistics.

Since SPSS is an industry standard, universities buy site licences and pay a lot to get them. They usually buy the yearly licences so they can upgrade pretty much as soon as a new version of SPSS ships. In other words, as an employee of higher education you have to repeat the entire mess once a year just to receive the mercy to continue looking at and working with your very own data.

At the University of Bern, the new SPSS licence code was a yearly ritual that got first announced in the group meeting, then it was passed down to a selected number of persons, who in return were allowed to give you the code if SPSS was considedered necessary for your research. Once you got the code you had to look for some EXE file in Windows Explorer to run an update program where you could enter the code. Then SPSS realized they could build the renew function into the Help menu tree and the whole ritual became sort of painless, despite still being rather silly.

Now life at the ETH is different. Since the Federal Government has tons of money, the ETH has site licences for pretty much every software package. You download it from a central server, copy the licence code from a attached textfile and within ten minutes you do what you actually want to do: getting some work done.

Of course, it can’t be that simple for SPSS 13. Attached to the installation program are two text files, one with an authentication code and one with a licence code (it takes a while to figure out the difference first). Of course, the given licence code doesn’t work and fails with some cryptic error message. However, even if it did work, the text file cheerfully informs me that it would only work until the end of April 2007 anyway.
Seeing this obstacle as a challenge rather than an annoyance we have the ability to use the authentication code in combination with the lock code to retrieve a new licence code from the internet. To my surprise, the connection to the SPSS server actually works and the program informs me that the changes will be applied once I restart the program. I close and restart the program - just to be greeted again with the notification that my licence has expired.

Repeat after me: proprietary software is bad for the customer. And use R instead of SPSS - it’s more powerful anway.

Update

Newton was so kind to forward two interesting links that compare R and other statistical software:

R, SPSS and SAS comparison (Englisch)
R and SPSS comparison (German)

By the way, if you’re in Switzerland and need support for R or would like to bring in your own ideas and resources, you may talk to the guys who have their offices one floor above mine. They support and mirror R: http://stat.ethz.ch/CRAN/