Minor refactoring

[KyrylR]

• This PR suggests a bug fix and I've added the necessary tests.
• This PR introduces a new feature and I've discussed the update in an Issue or with the team.
• This PR is just a minor change like a typo fix.

---

[KyrylR]

Link: https://cdla.dev/permissive-2-0/

[KyrylR]

Link: cdla.dev/permissive-2-0

This one is from LWK, and here is a AI summary of the differences:

Short answer: MIT-like in spirit, but not the MIT license. It is a permissive license designed specifically for data, not software code.

Main differences:

Topic	MIT License	CDLA Permissive 2.0
Primary target	Software	Data / datasets
Permission	Use, copy, modify, merge, publish, distribute, sublicense, sell software	Use, modify, and share data
Redistribution condition	Include copyright notice and MIT license text	Include the CDLA license text with shared data
Outputs/results	Not separately discussed	Explicitly says Results are unrestricted, including ML models and model insights
Warranty/liability	No warranty, no liability	Similar “as is” and liability disclaimer
Copyright notice	MIT requires preserving copyright notice	CDLA 2.0 requires making the agreement text available; it is less copyright-notice-centered
Sublicensing	MIT expressly allows sublicensing	CDLA focuses on sharing data under the agreement; results are unrestricted

The biggest practical difference is this: CDLA Permissive 2.0 lets you use a dataset, modify it, redistribute it with the license text, and use outputs/results without restrictions. That makes it broadly permissive like MIT, but it is more tailored to data and ML workflows.

For ML use: if you train a model on CDLA Permissive 2.0 data, the license explicitly says it does not impose restrictions or obligations on the resulting model or insights.

Sources: CDLA Permissive 2.0, MIT License - OSI

Not sure how it got into LWK, but at the same time it does not cover code at all, so yeah

[Arvolear]

LGTM!