Mondrian: Is It Heading Towards Closed-Source?

[yonghuapu]

Background

Mondrian is an OLAP (Online Analytical Processing) engine that plays a crucial role in the Pentaho BI suite. It has been a significant open-source project, enabling extensive data analysis capabilities for developers and organizations.

The Issue

Recently, while attempting to compile Mondrian from its master branch on GitHub, I encountered an issue. The POM file specifies a repository that requires authentication:
<repository> <id>pentaho-public</id> <name>Pentaho Public</name> <url>https://repo.orl.eng.hitachivantara.com/artifactory/pnt-mvn/</url> <releases> <enabled>true</enabled> <updatePolicy>daily</updatePolicy> </releases> <snapshots> <enabled>true</enabled> <updatePolicy>interval:15</updatePolicy> </snapshots> </repository>
Without access to this repository, the build process fails, making it impossible to compile the project locally.

Potential Explanations

This situation raises a critical question: Is Mondrian moving towards becoming closed-source? Historically, open-source projects have maintained publicly accessible repositories. The requirement for authentication could indicate a shift in strategy, possibly towards more controlled distribution.

Impact on the Community

If Mondrian indeed becomes closed-source, this would significantly impact the community:

• Developers: Those contributing to Mondrian would face barriers, reducing the number of community contributions.
• Users: Organizations relying on Mondrian might need to reconsider their data analysis tools.

Call for Transparency

I urge Pentaho and Hitachi Vantara to clarify their intentions regarding Mondrian's future. Transparency is crucial for the community to understand the direction and prepare accordingly.

Suggestions

In the meantime, developers and users might need to explore alternative open-source OLAP solutions or seek community-driven forks of Mondrian if it indeed becomes closed-source.

Conclusion

The open-source nature of Mondrian has been a significant strength, fostering innovation and collaboration. Clarification from the maintainers will help the community navigate these uncertainties.

[julianhyde]

This is a worrying direction for the Mondrian project. I agree that if it can no longer be easily built from source then it is effectively no longer open source. That said, it is not Hitachi’s responsibility to be sole maintainer of this code. If you - or several of you - think you can do better, you can make your own copy of the project. A fork is not a step to be taken lightly, but is an option to consider. The license - like every open source license - allows a fork. The discussion at this point should not be “Why is my free lunch no longer free?” but “How can we help?” Julian (Original author of Mondrian)

[yonghuapu]

Hi Julian,

Thank you for sharing your insights and concerns about the future direction of the Mondrian project. I completely agree with your points, especially regarding the importance of maintaining the project's true open-source nature by ensuring it can be easily built from source.

We, as part of the community, should step up and take responsibility if we want to see continued development and improvement.

Considering your suggestion about forking the project, I believe it's a step worth exploring. It might not be an easy decision, but if it ensures the project's vitality and aligns with the open-source principles, it could be the right path forward.

Thank you again for your guidance.

[julianhyde]

To be clear, forking is available as a last resort. First, we should offer PRs to fix the problems we see, and hopefully Hitachi/Ventara will accept them.

[yonghuapu]

Hi Julian,

I raised the question "Is Mondrian Heading Towards Closed-Source?" because I have some ideas I'd like to contribute via PRs, but I'm struggling to build the application locally in IntelliJ IDEA due to many dependencies being unavailable for download.

I've read some of your blogs from over a decade ago, which contain a lot of interesting thoughts about Mondrian, like how Mondrian should obtain statistics for tables and columns. I think this would be a great way to improve performance, but I haven't seen anyone implement these ideas in the latest versions of Mondrian.

I want to submit some PRs to address this, but I'm facing issues. Despite having ideas, I can't build the application locally to test and implement them. I'm currently using Mondrian 4 and have had to rely on googling to find some hard-to-find dependencies to build the application.

I've carefully studied some of Mondrian's source code, and it's a great design. I hope Mondrian continues to develop well.

Thank you for your time and looking forward to your response.