update README: Adding ToC, how to use the notebooks, descriptions. Little cleanup

[Urfoex]

epic &4
issue #35

[Urfoex]

Do we still need the WELCOME.md? @alxn4 @getml/all

[Urfoex]

Do we still need the WELCOME.md? @alxn4 @getml/all

Removed together with binder, as it is just partly working.

[liuzicheng1987]

Why are we using podman here? I think Docker is the standard here and using a tool that many people have never heard of will make adoption harder.

[Urfoex]

Why are we using podman here? I think Docker is the standard here and using a tool that many people have never heard of will make adoption harder.

Podman is a mostly hazel-free drop-in replacement for Docker. Because of its higher security, it is included in Redhat since version RHEL 8:

https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/8.0_release_notes/notable_changes_to_containers

https://www.heise.de/news/Red-Hat-Enterprise-Linux-8-freigeben-Mehr-Flexibilitaet-bei-Software-Versionen-4416725.html#nav_container_im__3

Not to mention: companies with more then 250 people and Mac and Windows users probably evaluated alternatives like Podman-Desktop and Rancher-Desktop because of:
"Commercial use of Docker Desktop at a company of more than 250 employees OR more than $10 million in annual revenue requires a paid subscription (Pro, Team, or Business)." (https://www.docker.com/products/docker-desktop/)

Define "many people".
https://github.com/containers/podman has 21K Stars.

While the docker open source part seems to be now moby
https://stackoverflow.com/questions/47793812/where-is-the-docker-source-code-located-github-repo
https://github.com/moby/moby
with 67K stars.

"Docker being the standard" for what?
The container world is much bigger.
https://en.wikipedia.org/wiki/List_of_cluster_management_software
https://en.wikipedia.org/wiki/OS-level_virtualization

For simple users trying the notebooks it would be nicer to have a full ready image on a registry.
It seems, there is a getml account;
https://hub.docker.com/u/getml
It would be nicer, to have a full image with notebook and engine on their and ready to run, instead of having the user to build it.
Or on another registry, like

• quay.io (redhat)
  

• gcr.io (google)
  

• public.ecr.aws (amazon)
  

• ghcr.io (github)
  

• ...
  

Or simply download the notebooks and run them locally inside their Jupyer-Lab.

Not to mention, podman is a drop-in replacement for docker.
In the repository, there is still just Dockerfile and docker-compose.yml. Just making sure, that the software and libraries inside are provided and run-able as user and not as root.
There is no extra adoption needed. Just reaching a wider audience.

[srnnkls]

Thank you @Urfoex. While I agree with everything you said, I would argue many people – in particular in the data science community – are not familiar with the nuances of Linux containerization. As docker is the best known implementation, we should use that throughout the examples. If you like, maybe, you can reference podman (like you have done in the root-level README).

[Urfoex]

Oh, in the docker/README.md. Sorry, I missed that part.
Adjusted.