DevOps plays a prominent role in mainstream technology and companies’ ability to deliver results. While the definition of DevOps encompasses the complete range of characteristics it covers, in practice, current tech culture tends to focus solely on the tooling aspect, neglecting the equally essential focus on the people and process critical in delivering value to end users. While cultures are difficult to change and redefine, a transformation is needed to fully capitalize on the potential DevOps has to accelerate software development.
Conway’s law–the concept that organizations ship software based on their organizational structure–outlines the need for visibility. This law revolves around the idea that in order for a software product to successfully operate, its developers must have established open pathways of communication to ensure each component works as a whole rather than as individual pieces. Unfortunately, today, many organizations still run into problems with how their teams operate when individuals own limited pieces of development.
Take a retail platform, for example. In the development of the platform, there will likely be several sub-teams that focus on individual components of the design. Perhaps there’s a team for the shopping cart service, another for the catalog service, a team for the underlying data platform and one for user account services. The list goes on. Typically, those teams do not have visibility into each other’s work and operate in their own silo. As soon as teams start thinking about ‘their’ component or ‘their’ microservice, a dangerous tunnel-vision takes hold. In some cases, there may be standup meetings across teams for coordination efforts, but this takes time, energy and steals focus from the important task: Delivering value to end users.
Alternatively, think about your favorite open source project. The inner workings were likely designed asynchronously, openly, remotely and at a global scale. Given the remote nature of open source innovation, developer teams don’t need to be co-located in the same workspace. While physically operating separately, open source developers have mastered alternative methods to operate with transparency in such a way that external contributions aren’t just welcome, but encouraged.
This same idea applies to proprietary software development, or InnerSource, a modern engineering approach that applies open source practices inside your organization to proprietary code. While a technological shift in practice, InnerSourcing is also a cultural revolution toward collaboration.
Companies of all types can benefit from implementing InnerSource practices. Some enterprise organizations that operate using InnerSource methodologies today include 3M, Ford, KPMG, Mercedes-Benz, Nubank, Spotify and more.
A critical component in making this change is prioritizing team visibility. If other teams in your organization are unable to see what you’re working on, you’re inherently (intentionally or not) creating silos and limiting efficient collaboration. In practice, there will be scenarios where company-wide visibility may not be feasible due to external regulations, highly secure environments or sensitive projects; however, these moments of limited project visibility should be the exception rather than the default.
When approaching InnerSource through an open source mindset, you will find that many DevOps principles still hold true. Here are a few strategies for implementing a collaborative open source culture when developing proprietary code:
- Use existing resources: When problem solving, using the existing materials available within your organization is a great place to start—it is possible someone has already tackled the issue you are addressing. If you find the right fit, you can reuse prior work and, if it has not been started, this is an opportunity to add value to your organization by creating a project that could help others in the future. When building upon someone’s work or starting a project, prioritizing visibility by providing an explanation of any technical requirements of the change will allow future project maintainers to determine whether it’s in the project scope, already a proposed change, on the backlog, or not a prioritized piece of work. This use of existing resources will help develop a collaborative culture over time.
- Documentation of work: Documentation is critical. Keeping track of the expected standards of the project and which tools or dependencies are required to build the project for local testing will help others in your organization know what to contribute. When in the early stages of contributing code to a project, documenting proposed changes without directly impacting the production version will allow you to receive feedback early on from the project maintainer about the direction you should take as they will be able to see your steps clearly outlined. To contribute code to a project, you can propose changes as a pull request, which doesn’t impact the production version but means you can get valuable feedback on the direction the project should take. Handling contributions by a pull request also ensures that branch protection rules apply and continuous integration practices are required.
While implementing these practices may not result in immediate changes, over time they will result in a significant cultural shift. And while cultural transformations require continued conscious efforts, embracing modern software engineering practices–learned from the open source community–will enable teams to break barriers for organizational collaboration.
