It's completely open and that's amazing. The ability to design plot.ly plots offline in a jupyter notebook and easily deploy it to our reporting website (Using Dash) is so quick and easy. Even the default interactive elements is plenty for most use cases.

The ReasonML group have done the same[1], and to great success IMO.

[1]: https://discord.gg/reasonml

Would you mind sharing that work-flow? I'm interested in how this is done, since i have failed to do it on several occasions.

I have failed to adopt it as well, and am still using Vagrant, but as my desktop is Linux, vagrant is a little bit of overkill.

The primary stopper for me is run vs start I think - and persistence of the most recent change I made in an environment.

`run` creates a container from an image and tries to start it. If it doesn't start you have a stopped container in `docker ps -a`. If it does start, you have a started container in `docker ps`. You could stop a started container if you wanted, and later start it again without having to use `run` all over again.

Sometimes running fails to start the container because there's a problem. You'll then have a stopped container.

What is in a typical make file for your projects?

It really depends of the project, but here are two examples, complete and simple: https://gist.github.com/bpierre/da5283f34a03b26e2833