Buggy

Currently my days at work are training and bug reporting. The first half of the day I spend reading about some aspect of the program and working through the examples that they have. The latter half of the day I spend entering in bugs that I find in the program, or making suggestions about how I think the program can be improved.

Well, last Friday I found a problem were the program was clearly not performing correctly, and this was not a small part either, this was like a major cornerstone of the program that was not working right. At first I was excited, because I found something very important while being at work for only a month. Then I kinda got scared, because if this was actually happening the program has serious issues. It was kinda like discovering a highly communicable deadly disease that eventually will kill everyone, it is a very important discovery -and now we're all going to die.

So I brought the issue up with Danny (programmer/statistician extraordinaire) and asked him what he thought, he got worried about it and had me show him exactly what was going on. So I showed him and sure enough the results weren't the expected results. We ran through all sorts of combinations and possible scenarios about why the program was failing. We couldn't come up with what was going wrong, so he decided to head back up and try to figure it out. As I was closing down the analysis he noticed that I had a check box selected that shouldn't had been way back on an early dialog box.

Dang it. I selected the box early in the morning when I was working with some other stuff and never unselected it because I had just been working with the same analysis all day, and then when I got sidetracked I forgot about it. That box meant that a whole different batch of assumptions were made and the program was actually performing quite well. Call off the CDC that deadly virus was just some soda I spilled on the microscope slide.

Grumble, grumble.