The Python Software Foundation (PSF) extended a helping hand to the packaging community at the recent PyCon US Packaging Summit. In an open forum, the PSF inquired how they could best support our ongoing efforts to improve the packaging ecosystem.
One proposal resonated strongly with attendees: the creation of a dedicated governance system for the packaging community, perhaps modeled after the successful Steering Council that oversees the development of Python’s core. This sparked enthusiastic discussion, with many recognizing the potential benefits of a similar structure for guiding and coordinating packaging-related initiatives.
The resounding support for this idea demonstrates a clear desire within the community for greater organization and collaboration. A formal governance system could provide a platform for addressing challenges, setting priorities, and driving innovation in a more coordinated and efficient manner.
The PSF has expressed their willingness to assist in establishing such a system, should the community decide to pursue it. This presents a unique opportunity to shape the future of Python packaging. Let’s seize this momentum and start a conversation about how a governance system could empower the packaging community to achieve its goals. Share your thoughts, ideas, and concerns. Together, we can build a stronger foundation for the future of Python packaging.
In my humble opinion, such governing council could help us break the status quo, where we only tend to accept proposals that have outstanding support, and we tend to shy away from making more controversial decisions, such as what should be a good enough default entry point to packaging for new users to the language. This is mostly because we don’t want to give the impression that we bless a given tool, even though this is what most users would like when they are first learning the language (nowadays all our documentation mostly just say pick one of these six choices, for almost every facet of packaging, which is a really bad experience for someone just wanting to tip their toes into packaging). The fact that pyOpenSci Python Package Guide — Python Packaging Guide (presented at the conference) is not under packaging.python.org is a good example of this.
Personally I would like to be in a place where we are more comfortable to accept proposals on a provisional basis, that work for most of the cases, even if on day one they don’t solve qll of the problems. Another good example of this is our struggle to try to standardize a lock file for the language.
Creating a better council based governance model would also provide a platform for all the projects that don’t live on there the PyPa org, that likely should have a say in where we want to push the packaging story (such as various core developers, conda, pixie, uv, poetry, Nvidia, ml and so on). This would also remove the impression that the decision is with PyPa rather than the wider community.