Now that we finished automating as much of the process as we could, it was time for data collection. Fastscan, the commercial
program we used to detect faults in circuits, gave us data that was difficult to read, so I wrote a program to parse the data
and compile it into a readable format that tells us which faults were detected for which pattern and how many times. However,
after double checking the first run of this program, we found a discrepancy in how many faults were detected. We spent an unfortunate
amount of time retracing the automated process trying to find the problem, only to find that my original program that generates randomized test
patterns was not randomizing the amount of times specified at input, but printing out only one randomized string that amount of times.
Fixing that problem only gave us more accurate Fastscan results and the discrepancy remained.
After painstakingly going through the data we found that Fastscan wasn't consistent with the data it gave us, and the program we had that
compiled a fault dictionary didn't handle Fastscan's inconsistencies. We are in the midst of editing the dictionary program to handle this problem.