Skip to content

6.1. License

What is a software license?

A software license is a legal agreement that details the conditions under which a software product can be accessed and used. This agreement is essential for defining the rights and responsibilities of both the developers and the users. It covers aspects like software modification, sharing, and usage in other projects, thus protecting the intellectual property of the creators and the legal rights of the users.

Software licenses are typically categorized into two main types: open-source and proprietary. Open-source licenses generally allow the software to be freely used, modified, and shared, while proprietary licenses impose restrictions on these activities.

Why do you need a software license?

Implementing a software license is crucial for several reasons:

  • Legal Clarity: Establishes a legal framework that dictates how the software can be used, preventing disputes.
  • Copyright Protection: Protects your intellectual property from unauthorized use or distribution.
  • Control Over Distribution: Dictates the terms of how your software is distributed, whether it is sold, given away, or integrated into other products.
  • Open Source Integrity: In open-source projects, a clear license preserves the ethos of free use and collaborative development, and governs contributions.

In an organizational setting, it's vital to choose a license that aligns with institutional policies and objectives. Consultation with legal and management teams is recommended to ensure compliance with these policies.

Example: MIT vs GPLv3

The MIT License and the GNU General Public License v3 (GPLv3) are two prevalent open-source licenses that differ significantly:

  • MIT License: Highly permissive, allowing almost unrestricted use, including in proprietary projects, provided the original copyright and license notice are included in copies of the software.
  • GNU General Public License v3 (GPLv3): This copyleft license requires that any derivatives of the software also be distributed under the same license, ensuring the preservation of usage, modification, and sharing rights.

Case Study: ElasticSearch vs OpenSearch

ElasticSearch switched from the very permissive Apache 2.0 license to a dual-license model due to concerns about cloud providers using their software without contributing back. This led Amazon Web Services (AWS) to fork ElasticSearch, creating OpenSearch under the Apache 2.0 license to maintain its openness and community contribution. This case underscores the strategic role of licensing in software development and community engagement.

How to choose your software license?

Choosing the right software license involves careful consideration of your project’s objectives and the legal landscape:

  • For Open Source Projects: Utilize tools like ChooseALicense.com to select a license that fits your preferences regarding how your software is used and shared.
  • For Proprietary Project: Work closely with your company’s legal team to ensure the license aligns with business strategies and regulatory requirements.

How to define the license for your project?

To implement a license in your project:

  1. Create a LICENSE.txt file in the root directory.
  2. Include the Full License Text: Ensure the complete, unaltered text of the license is included to make it legally binding.

You can find open-source license text on this website: https://opensource.org/license.

Should you choose a different license for AI/ML models?

Licensing AI and ML models necessitates addressing unique considerations related to the nature of data usage, model training, and deployment:

  • Data Privacy and Usage: It's crucial to ensure that the licensing agreement covers the handling of potentially sensitive or personal information in compliance with applicable privacy laws, such as GDPR.
  • Model Reproducibility: The license should specify if and how the model can be used to generate derivative models or reproduce results.
  • Commercial Use: You need to decide whether the model can be used for commercial purposes. Some licenses may restrict commercial use to promote free academic and research utilization.
  • Attribution Requirements: If required, the license should mandate users to credit the model creators or reference the original research.

Here is a list of some popular licenses used for AI/ML models that address these considerations:

  • [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0: Permits almost unrestricted use, including commercial use, while still requiring attribution.
  • [GNU General Public License (GPL) v3.0](https://en.wikipedia.org/wiki/GNU_General_Public_License: Ensures that derivative works are distributed under the same license terms, suitable for models where sharing improvements is desired.
  • [MIT License](https://en.wikipedia.org/wiki/MIT_License: A permissive license that allows almost any use with minimal restrictions, provided that the license and copyright notice are included.
  • [Creative Commons (CC BY 4.0)](https://creativecommons.org/: Suitable for datasets rather than models, requiring attribution but allowing for commercial and derivative uses.

How should you integrate the license of your software dependencies?

Effectively managing the licenses of software dependencies ensures legal compliance and supports community and user trust. Here are streamlined steps for integrating these licenses:

  1. Inventory Dependencies: Use tools like SPDX to list and document all third-party components and their licenses.
  2. Review License Terms: Analyze what is permitted or restricted under each license.
  3. Assess Compatibility: Check for conflicts between dependency licenses and your project’s intended use.
  4. Automate License Compliance: Employ tools like WhiteSource or Black Duck to automate license management, integrating them into your CI/CD pipeline.

These steps help maintain legal integrity while leveraging the benefits of diverse software components.

License additional resources