I was wondering if anyone had to design a data base for "Double Data Entry" for data accuracy purposes.

If so what were some design considerations and challenges to create the solution. I'm interested in how the 2 sets of data were entered, and what techniques were used to validate the data.

Thanks in advance.

I've used blind double entry many times (but not for past 6-7 years). Blind (or double entry) is important to guarantee 100% accuracy of data. It usually means two different people enter the same information. This is because, particularly when reading written material, two people can interpret differently. But it can also mean same person enters a set of data and then enters same set (in same exact sequence) again.

Regardless of the business rule behind it, the data-entry would be the same. Envision: Load tickets: These have a ticket# and hold lines – which are logs on a log truck which is written down by a log scaler. These tickets need to be input. Once verified (double entered), the data would be imported into a real data file for payment.

The first time I created a blind double enter in FileMaker was using vs. 6 and I created two sets of records. However, if I needed to perform same function today, I would use repeating fields (2 reps) for each time a line is entered (see attached).

I didn't get fancy here ... there are many protections I would build in and buttons so a User would be isolated from the other set and so data wouldn't be imported into real tables until 100% accuracy was met. Anyway, I hope it is helpful.

BTW, sequence is very important here so tickets must be kept in proper order. When young, and working in mainframe, I assisted once in double entry. I dropped the stack of tickets. I wasn't asked to help any more and was sent back to mainframe operations. :crazy2:

BlindEntry.zip