Plainionist Plainionic Programmer

How to implement the Clean Architecture?

Did you enjoy reading Uncle Bob’s Clean Architecture?

Book: Clean Architecture

I did! As with every book from Uncle Bob’s it is motivating and inspiring, right?

So let’s take his ideas and realize these in our projects to gain what he is promising!

But how do I start? How do I transform an existing code base - following a layered (web) architecture - into the Clean Architecture?

The Clean Architecture consists of multiple layers organized as circles while dependencies are only allowed from outer circles to inner circles. The inner circles contain the business logic. All details, devices and frameworks are in the outer circles.

In this blog series I will share with you the questions I had and the experiences I made when implementing the Clean Architecture.

When I read about the Clean Architecture one of my projects already had an existing code base and has even been released several times. Its quality was basically OK but there was clearly improvement potential. So I decided that this would be a perfect pick for my experiment …

The project

The project of choice is an “agile backlog visualizer” - I will call it Athena in the context of this series ;-)

Athena’s mission is to bring maximum transparency into the backlog of our agile teams. It fetches all work items from our Team Foundation Server (TFS) and visualizes those nicely on an intra-net page to give us the views and reports we need …

The Backlog

Our backlog is ranked and has multiple cut lines due to usage of confidence levels. The cut lines are visualized with the background color of the row. The darker the color gets the less likely the feature can be delivered for this release.

Athena shows our ranked backlog in a simple table with some filters. It also applies the capacity of our teams to illustrate different cut-lines in different colors.

You may wonder why we have fruits in our backlog … this is a different story which I may tell the other day ;-)

Work balance

Having multiple teams working on the backlog we are looking for a balance of the remaining work compared to the teams availability across all teams.

As we have multiple teams working on that backlog we want to know how the different teams are loaded.

The Burndown

Our burn downs visualze the burned efforts as well as the projected burn down based on the past velocity. We have 'touch down lines' which show how work packages got re-scoped during the development.

And of course we want to track our remaining work using a burn down.


We also have quite some conventions for our backlog, e.g.:

  • each work item must have a parent (except the roots of course)
  • each work item must have certain category tagging
  • iteration paths within parent-child relationship have to be consistent

So we also have a page reporting violations to these conventions.

Tags: clean-architecture