Add user interface to help beginners with creation of suitable configuration files

[nuest]

At the eLife Sprint 2019 @innovationchef wanted to develop a user interface for generating Dockerfiles. We (@nuest @sje30) suggested to instead make it a user interface to create suitable configuration files that could be picked up by repo2docker. A first prototype can be found at https://github.com/innovationchef/easy-docker/ (just run the container :-) ). GIF:

I suggest to make such a wizard user-interface part of BinderHub (e.g. https://mybinder.org/config-generator) and not part of repo2docker as discussed here (with some useful ideas!), because users who need a UI for this task are unlikely to be able to run repo2docker locally.

@innovationchef Please keep this issue updated with you plans. It would be great if you could extend the user-facing workflow a little bit and we can leverage your prototyping to p

---

The prototype is based on Java, so not easily integrated with BinderHub, but hopefully it allows us to iterate through some variations. IMO...

• this should be possible with a plain JavaScript UI
• the dev should "wait" for progress on the awesomebar (Awesome bar/landing page redesign #844) before being implemented
• a first iteration should simply provide a download of an archive with the generated files, a later version might analyse a code repository and create a PR 🚀

[betatim]

One thing on "tech": it feels like several people at the team meeting last week were Ok or enthusiastic to use material design as the "UI toolkit" for a newer BinderHub UI. As a general goal we want all the pages to "feel the same" so I wanted to leave a note here about a likely future choice for the rest of the pages.

[nuest]

Forgot one thing: @minrk mentioned that such a UI might also reduce the need for specific stacks (cf. https://github.com/jupyter/docker-stacks) because users can more easily create the environment they need via configuration files, and not by selecting a variation of an image in the Jupyter Stack list.