Design, document and improve the license and license rules contribution processes

[pombredanne]

Context

The license and license rules contribution processes is not working efficiently.

Today this is based on a combo of

• issues entered here and in the licensedb static repo
• PR here and entries in DejaCode license app
• sync of the license DB between here and DejaCode
• sync of the license DB with SPDX new releases

Problem

Also:

• There are over 200+ licenses issues pending
• the process is not efficient and slow
• it is not easy for contributors to contribute

Solution

We need to design, document and improve the license and license rules contribution processes.

Some areas to consider:

• Add requirement for test files for each license rule #4953
• RFC: allow adding more data to YAML frontmatter - syntax and proposed fields #5110
• move the licenses and rules in their own repo
• make the DejaCode license app accessible to any vetted contributor
• Combine efforts with OSI

[mjherzog]

There is already a draft document at: https://docs.google.com/document/d/1mLNituphqNwaOglUdn2FujHl4mRrBky8/edit?usp=sharing&ouid=109207805937002014344&rtpof=true&sd=true (anyone with link can access)

[armijnhemel]

There is already a draft document at: https://docs.google.com/document/d/1mLNituphqNwaOglUdn2FujHl4mRrBky8/edit?usp=sharing&ouid=109207805937002014344&rtpof=true&sd=true (anyone with link can access)

This is very useful for AboutCode team members (and to understand why certain decisions are made), but the problem with the current (and described) approach is that it doesn't scale. As @pombredanne said there are quite a few open licenses, PRs and other issues that need to be tackled and that are currently open. It is also not ideal for third party contributors.

[hesa]

I propose, as a short term solution, to update the github issue (template):

• add checkboxes in the issue (template), see below
• set priority to low - allow the reporter to raise it
• ask for references - e.g. if the user states the license is a Source Available, the user can copy/paste the relevant text
• ask for license usage - where is the license used?, how many or which projects use the license?, ...

With the checkboxes the issue reporter/author can give hints about the license, hopefully speeding up the process of approving the license.

Examples of such checkboxes:

• ML/AI
• Source Available
• Permissive
• Weak Copyleft (copyleft limited)
• Strong Copyleft (Copyleft)
  ... perhaps the last three should be in a radio button group instead

What do you think?

[armijnhemel]

What do you think?

Anything that makes the process easier is worth doing.

[mjherzog]

This is a reasonable low-cost experiment