Rug is the backbone of Atomist’s features. Atomist is about automating away all the distractions from writing and operating great software. Rug provides the tools and infrastructure to make complete automation a reality. The model that underpins Rug helps you automate common tasks and react to changes in your development ecosystem.
The Rug ecosystem includes a programming model, runtime, test runner, and package manager. The Rug programming model is expressed as TypeScript module interfaces and classes. The Rug runtime runs as a service, accessible from any Slack that has invited the Atomist Bot (try it in Atomist Community Slack). There’s also the Rug CLI for local use, essential for Rug development.
Rug is a medium for code that modifies code. Rug helps developers automate development. When a coding task is common, tedious, nit picky, or hard to remember how to do correctly, there is value in encoding how it’s done, instead of performing the typing every time. Not only does the automation reduce mistakes, it serves as documentation for the process.
Automating your development tasks¶
It is said that good developers are lazy and like to automate their work. However, the tools to drive that automation have been somewhat sparse and crude when we consider the sheer complexity of projects nowadays.
Triggering an automated response¶
In some cases, a response to a system event should be automated so the team can focus on the things that require human attention.
In Rug, this is achieved through event handlers.
Triggering a human decision¶
Automation is fantastic but humans are the sole judges. Atomist gives you the power to implement new skills that can be triggered by a team member at when needed.
In Rug, this is achieved through command handlers.
Creating new projects¶
In a world of rapidly evolving software, creating new projects has become a task performed much more often than in the past. Meanwhile, the complexity of projects has grown dramatically with configuration required for logging, CI, dependency management…
It appears clear that automating the generation of projects is a prime for any team willing to move fast but with repeatable quality.
In Rug, this is achieved through generators.
Automating the creation of projects is a great step forward but it cannot stop there. There are tasks that are repeated on a daily basis and doing them manually can be error prone, not to mention rather boring. Let’s not forget that code quickly becomes legacy that nobody knows really about any longer.
Automating those changes is an asset for any developer who wishes to focus on delivering great software without wasting time in mundane tasks.
In Rug, this is achieved through editors.
What does using Rug in your team look like in practice? Here are just a few examples.
- Helping technical leads to guide development teams in best practices on various technologies from initial project creation through to the full lifecycle of a project
- Safely applying and evaluating new technologies to existing projects
- Helping open source project owners to guide their users on how to start out with, and continuously update and evolve, the software based on their work.
- Helping to apply best-practice tools and techniques from the microservices toolbox
In recent years, the DevOps trend has shown us that concerns about software does not stop once it has been delivered. Software exists thanks to those who designed and developed it but thrives thanks to those who operate it. At Atomist, we believe those two sides live in the same world and more must be done to unite them. Atomist brings everything together through event-driven development.
A common setup today is as follows:
- A project’s source code lives in GitHub
- A project is automatically built and tested in a CI service
- A project is usually automatically delivered in a forge somewhere
- A project may even be deployed automatically in an environment
- A project is then operated, monitored, and cared for in that environment for users to enjoy
- Issues are created
During all those phases, a massive amount of events were triggered: a commit was pushed, a build succeeded or failed, the project was deployed, the service failed in production…
Atomist believes that all these events bring all the team members as one. However, not all events may not be able relevant to a team at a given time. Moreover, it seems appropriate to think that we should also automate the response to some of those events. This is why the Rug programming model has a holistic view of development and operation, allowing automation or user intervention at every step.