Partial Solution of Bank OCR Kata

http://codingdojo.org/kata/BankOCR/

Overview

This is a partial solution to the Bank OCR Kata in two languages: Ruby and Javascript. The goal for the Ruby solution was to see how far I could get in ~2 hours, documenting the steps I would take towards a complete solution. (See Status below.)

I Later ported the solution to JavaScript to see how closely I could reproduce the solution, including Ruby structure and coding style, in JavaScript.

There is no runtime environment at this point. The Ruby code is exercised via the RSpec tests. The JavaScript code is exercised by the Jasmine tests.

Ruby Solution

Steps to run the code

This code requires a standard Ruby development environment. The basic components are:

• Ruby 2.1 (or newer).
• Build tools. XCode or GCC and development libraries. Needed to build gems.
• Bundler 1.12 (or newer). This can be installed with gem install bundler

Start by installing necessary gems

cd [folder where you cloned the code]/ruby
bundle

Then run the tests

bin/rspec

Or use guard to run the tests automatically as you work

bin/guard

Javascript Solution

Steps to run the code

No installation is necessary. Just open the Jasmine test runner in a browser: javascript/SpecRunner.html

Status

I completed User Story #1 and #2. While not directly testing the user stories, number_parser_spec.rb contains specs that exercise the functionality of #1 and #2. There are probably some edge cases I'm not handling. For example what happens if the string passed to NumberParser isn't exactly 27 characters? I think that will cause an exception currently. Of course, this could be mitigated in a higher layer as well. FileParser could be written to always parse out exactly 27 character strings. If it were a client project, I would spend more time exploring the edge cases before moving on to #3

User Story #3 should be straightforward. It will require completing and wiring together Machine and FileParser, and possibly a FileWriter class.

User Story #4 will require adding functionality to DigitParser that can produce all possible digits for a given string. Then NumberParser will need to work through combinations of digits, checking each for a validity and storing the possible values. Then Machine will need to work out what do to from there with help of FileWriter. I haven't given this much thought beyond that.