These AMA’s seem to be taking off - nice idea, thanks Brett! Here is mine, happy to answer any questions anyone may have
You can see my nomination statement here. I’ll copy the “what would I focus on” part of it here:
If elected to the Steering Council, my priorities are threefold. First, to help establish the council on a transparent, predictable path while improving community review processes and decision-making workflows. Second, to improve support for native code, compiled dependencies, and hardware needs across scientific computing, ML/AI, and performance-critical workloads. Third, to strengthen the interaction with and support for distribution packaging, spanning Linux distributions, Conda, Spack, Nix, and beyond.
Thank you for stepping up to serve on the PPC. Your background with NumPy and SciPy brings a crucial perspective. Given your deep roots in the Linux ecosystems, can you also commit to actively advocating for the needs of Windows users? Since Windows developers and end-users represent the majority of the Python user base, often facing problems with compiled dependencies and a lack of bundled compiler toolchains, can you promise to keep them in mind as well?
The PPC will take on an important large amount of work. How much weekly time can you commit to for the in-depth review of proposed PEPs that don’t touch your preferred interests?
Regarding advocating for the needs of Windows users and keeping them in mind: definitely! I am fully aware that we have a lot of Windows users, and that building from source is significantly more challenging for them than it is on most other platforms.
As a recent example of prioritizing the needs of Windows users: I spent effort to get NumPy and SciPy wheels for win_arm64 shipped and chose to make it a Tier 1 platform for NumPy despite there not being much demand yet - any other (non-Windows) platform wouldn’t have gotten that status with similar usage numbers, win_arm64 got it precisely because building from source is more challenging.
I’d also say that I don’t specifically have deep roots in Linux ecosystems. It’s been my preferred platform to develop on for a long time, however I have also used Windows for years, both when I was in academia - where it was typically installed on shared machines in the physics labs I worked in - and in jobs in industry. My current dev setup happens to be macOS-first, using Linux only in containers and in case I need a GPU-enabled remote dev box. And the two distros I listed as personally contributing to (conda-forge and Spack) both support Windows; that they work cross-platform is part of their appeal.
Good question. The time commitment that being part of the inaugural steering council (SC) was something I weighed carefully before deciding to stick my hand up. There isn’t any more time in the week that I can take away from family, friends and time spent away from a computer. So, if I do get elected, I need to drop or significantly reduce my time on day job or volunteer open source commitments.
I decided to spend at least 5 hours a week on SC-specific duties, and that those 5 hours should be additional to my existing work on Python packaging topics. There’ll be some responsibilities at Quansight I can hand over or delegate - and given that the SC starting to function well is important to the whole community, and hence well-aligned with Quansight’s mission, I am in the lucky position that I can do that. However, I will also have less bandwidth for personal open source contributions - and most likely that means less time spent on SciPy[1].
I’ll note though that I expect that a lot of the extra SC time, especially in the beginning, will be spent on getting the council up and running, improve processes for PEP review and feedback gathering, and reporting back from the SC to the community. A main benefit of having a five-person council is that responsibilities for PEP review can be shared. Each PEP requires the attention of more than one council member I’d think, but I wouldn’t expect that I’d spend most of my extra SC time on PEP review of topics that I’m least knowledgeable about (that wouldn’t be a very effective way of working).
Both because that is the project I spend more time on than on anything else at the moment, and because it has got an excellent group of maintainers I trust, so it’s not like things will stall or go off the rails if I’m less available there. ↩︎
