hrs / bank-ocr-kata

A partial implementation (steps 1 and 2) of the "Bank OCR" kata.

What should I do with this?

This is just a library, so there isn't anything executable in here. You can, however, run the tests:

$ rspec

Some things I might do in the future, with a bit more time:

For user story #3:

• Seems like it's about time to create an executable script. It should take in one or more file paths and print the resulting findings.
• Since we're dealing with multiple files, we might want to create an object that encapsulates a collection of entries corresponding to the contents of one file. Not sure about the right name in the problem domain... call it a Page, maybe? The stuff to be written might be Page#to_s.
• We created an IllegibleDigit earlier on, so we can tack a #to_s method onto it and handle the ? cases without too much trouble.

For user story #4:

• We'll need to examine each digit in each invalid entry. The brute-force solution would be to generate every block of characters that's one error away from the existing digit, then search that space for digits that produce a working checksum. That's, what, 3 ^ 9 possibilities for each digit? That's only about 20,000, which seems brute-force-able.
• Each digit's parsed from a block of characters. I'd been keeping that block as state in each of the Digit objects (and in each IllegibleDigit), but I ditched it along the way since it wasn't useful. I'd probably add it back now!
• Ambiguity is a bit tricky. We'll need to either iterate through the digits and maintain a flag or recurse though 'em with a conditional.