asa-holland / gildedrosekata

Exploration into a classic refactoring kata as a way to learn C#.

Gilded Rose Kata with C# Core

What's this kata thing? Code katas are concise exercises that focus on principles of design patterns and refactoring in software engineering. They serve as hands-on training grounds where developers practice applying these principles to real-world scenarios. By working on katas, software engineers learn how to recognize and implement design patterns effectively and refine their refactoring skills.

Why the Gilded Rose? This repo is branched off a template created from Emily Bache's GildedRose-Refactoring-Kata repository, which in turn is derivative of Terry Hughes original work. You can also read Bobby Johnson's description of the origin of the kata. I picked this kata to try specifically because a) I had seen it repeatedly referenced across the internet and b) as a kid I used to play World of Warcraft, from which the 'requirements' have been derived. Who knows, its possible that my long-abandoned WoW character could be in the trade district of Stormwind, drinking a pint at the Gilded Rose Inn as you read this very sentence...

Why C# Core? I'd never used C# before this Kata, and wanted to learn after I read Bobby Johnson's article "Why Most Solutions to Gilded Rose Miss The Bigger Picture", which suggests better practice at handling a legacy code situation if the Kata is performed in the original C#. After some desk research, it seemed to me at the time that most C# applications were done in tandem with the .NET core framework - not written using C# in isolation, so I chose to use C# within the .NET framework (aka C# Core).

On that note, here are the Gilded Rose Kata Requirements.

Changes from Emily Bache's Refactoring Kata

The following changes were made before starting:

• all non-English requirements.txt files were removed because sadly I am only fluent in English
• all non-C# Core projects (including standard C# and 'C# Verify') were removed
• the N-Unit style [Fact] based Approval Tests file was removed, while the X-Unit style tests were kept - this was because I am more familiar with X-Unit style test suites like Pytest for Python and Jest for Typescript, and based on my limited desk research it appears that N-unit style frameworks are less common and less in style. Also, since both the N-Unit and X-Unit test suites in Emaily Bache's repostiory were failing (IE, this follows Michael Feathers definition of a 'legacy system' - one with no passing tests), the kata is starting from square zero.
• the GildedRoseRequirements.txt file was converted to Markdown, because I prefer reading text files in Markdown using VS Code's live preview functionalities

Setting up the Kata

I ran the Kata inside Visual Studio Code (my preferred IDE) on a Windows 11 PC. At first I thought I needed to use Visual Studio and struggled to configure equivalent of the Conventional Commits extension there before just trying the .NET build on VSCode - lo and behold, works just fine. The following steps were taken to get things up and running:

• Install the .NET SDK (https://dotnet.microsoft.com/en-us/download/dotnet/thank-you/sdk-7.0.401-windows-x64-installer?journey=vs-code)
• Add the .NET SDK installation directory to Path (on Windows)
• Install the .Net Framework (https://dotnet.microsoft.com/en-us/download/dotnet-framework/thank-you/net481-web-installer)
• run cd csharpcore
• run dotnet build to check the build and ensure eveything runs correctly
• run dotnet test to run the unit test suite

These last three commands have been abstracted into a bash script, which can be executed from the root with bash main.sh.

You can also run just the unit tests or just the acceptance tests using bash unit.sh or bash acceptance.py respectively.

CodeSense Score Report

The following scores are intended to be used to document the code health before and after the refactoring kata was performed:

Metric	Details	Before	After
Overall Code Health Score	---	8.35	10
Bumpy Road Ahead*	Bump Count	5	not listed
Deep, Nested Complexity*	Nesting Depth (conditionals)	6	not listed
Complex Method*	Code Complexity	19	not listed
Large Method*	LoC	71	not listed

*Note: these indicate the highest score warning (least desirable) on any method in the system under test. Ideally the metric is not listed.

To read more about CodeSense, check out :