Fudging data

It's late, and I'm running tests on code for a client.

The code's been driving me crazy.

Instead of searching and shaping existing data to test my algorithms, I'm adding some code simulating data -- fudge. As I control the data, I can look for a clear, pre-determined outcome. If I keep getting what I expect, then I know the algorithms work. Then, I'll comment out the code, and will let the real data go through the algorithms.

Of course, if I was a climate scientist, this would be all evidence of improper tampering of results.

See Tim Lambert over at Deltoid for more.