Get YARP-enabled workspaces on the cloud with a single click!

[pattacini]

We've got exciting news we'd like to share with you today ✨

There's now a super quick way to start experimenting with our ecosystem without the hassle of installing any software components. You'll need only the browser! 🚀

Is this a kind of magic? 🧙‍

Not at all! Infrastructure as a Service (IaaS) have been catching on for a while by then, making available lots of resources online straight from your browser 🌐

Just click on this Codenvy factory link (you'll be prompted to sign up - it's free!) and a fresh full-fledged Ubuntu system plus IDE will be ready for you on the cloud with YARP and iCub suites pre-installed 😲

We chose Codenvy because it is a versatile fully free environment with a swift and gentle learning curve 💹

What's more? You can easily share your workspace with a peer!

What about graphics? Simply install a VNC client – RealVNC is a nice one – to get a view on your cloud desktop to play with the iCub simulator.

This is not the whole story ☝

Thanks to the Docker engine we are employing under the hood, you can also rely on different IaaS providers such as Gitpod ⚙
Configuring Gitpod is like a walk in the park:

• Drop a file called .gitpod.yml in the root directory of your GitHub repository hosting some fancy code you aim to test. The file needs to contain a pointer to one of our Docker images. For example:

  image: pattacini/technical-evaluation-stack:gitpod

• Click on https://gitpod.io/#https://github.com//your-repository and you're done! 🎉

Gitpod certainly looks very modern and slick (it is based on Eclipse Theia) but doesn't allow for sudo at run-time (could be useful to customize the workspace without tinkering with Docker). Gitpod is also free only for 100 hours/month of usage and for public repositories, although the latter hitch can be easily worked around.

Further, when you share a running workspace you'll share also your access to GitHub, which might not be desirable ⚠

Closing remarks

This infrastructure represents only a first step, but should already be able to ease the time-consuming processes underlying the preparation of training courses.

Both Codenvy and Gipod have pros 👍 and cons 👎 obviously. However, it's important to warn you not to expect to put your hands on quantum computers 😉
After all, these services are free of charge!

We'd love to hear your feedback ❤

[pattacini]

Following up the discussion we had today with @traversaro, providing a Docker image attached to some public repository through Gitpod would be a really nice addition to a scientific paper in our field toward the direction of reproducibility of results.

Ideally, we should:

1. Provide the code accompanying the article in a GitHub repository
2. Tag the release with Zenodo
3. Prepare a Docker image with everything needed to run the experiment and publish it on DockerHub
4. Drop in the GitHub repo the .gitpod.yml pointing to the Docker image. You can tell Gitpod to download the correct version of the code within its environment.
5. Include the instructions on how to run the experiment in the main README.md

Gitpod here has some advantages over Codenvy since it sports more RAM and powerful CPU so that for example a resource hog like Gazebo should be running fine. In this case, we don't share anything, thus the caveat above on Gitpod doesn't apply.

[pattacini]

By installing noVNC within the Docker image, we no longer need an external VNC viewer! 🎉