---
tags: Scientific Python
---

# Scientific Python Ecosystem Coordination Birds-of-a-feather

[This document](https://hackmd.io/@tupui/scientific-python-bof)
[Slides](TODO)

[Scientific Python Communit Discord](https://discord.gg/vur45CbwMz)
[Project homepage](https://scientific-python.org)
[SPECs](https://scientific-python.org/specs/)

Attending:
- Stéfan van der Walt @stefanv
- Jarrod Millman @jarrodmillman
- Pamphile Roy @tupui
- Juanita Gomez @juanis2112
- Sebastian Berg @seberg
- R Ely @ohrely
- Brigitta Sipőcz @bsipocz
- Charles Harris
- Greg Lee
- Matthias Bussonnier @carreau
- Juan Luis Cano @juanlu
- Paul Ivanov @ivanov
- Nezar Abdennur @nvictus
- Matthew Feickert @matthewfeickert
- Henry Schreiner
- Jim Pivarski
- Jon Guyer @guyer
- Dan Schult @dschult
- Chris Sewell @chrisjsewell
- Dario Taraborelli @readermeter
- Jason Grout @jasongrout

... and quite a few other community members.

We will try to keep this session interactive.

> There is **community outreach tonight at 6:30, and two talks tomorrow**:
> - 11:30am Github -> tiktok
> - 1:15pm Maintainers track.

stefanv: More formal effort to coordinate efforts across the scientific Python ecosystem.

Jarod: Bof same size as early scipy in Cal tech and amazing community.

- SP: we do specs: guidelines to help mainainer to align on things.
- common ci, lazy load imports,
- hugo theme to have same appearance it complements the Pydata sphinx theme,
- [Central place for all the calenders](https://scientific-python.org/calendars/)
- Your project can join this with a PR ([example PR](https://github.com/scientific-python/scientific-python.org/pull/251))
- Matthias thinks: Jupyter calendar should move there. If someone want to help...

- Analytics:
- There is a big pushback agaisnt google analytics in Europe, importatnt to have analytics w/o targetting. tracking users.
- Providing a server, self-hosted based on plausible.io,
- 

- Blog: Recently wrote a blog and we are looking for writers. designers ...
write a tutorial for the community, for example
- https://blog.scientific-python.org

- NetworkX have the GSoc student write on the blog

- All matplotlib content is there now.

- [Discuss](https://discuss.scientific-python.org/) and discord for async and sync communication.

- [Devstats](https://devstats.scientific-python.org/)
- Still growing (more statistics), but a few dashboards exists
- Hopefully useful for grant proposal or progress reports when the grant is running
- Scripts are all there to modify it yourself

- Matthias Notes: make me think of [napari #3886: Add issue and PR drawdown tracking page](https://github.com/napari/napari/issues/3886)

# Discussion

https://discuss.scientific-python.org

> Q: Is everyone in our community on git and GitHub?

- figure out how to get stuff from gitlab or sourceforge, if necessary
- grpk is developed on sourceforge, but may be mirrored on Github

Jarrod: all the stuff is prototype-y
@jarrod Complete notes here

> Q: Can you share an example of a SPEC?

https://scientific-python.org/specs/

- [SPEC 0 — Minimum Supported Versions](https://scientific-python.org/specs/spec-0000/) ([aka NEP 29](https://numpy.org/neps/nep-0029-deprecation_policy.html))
- [SPEC 1 — Lazy Loading for Submodules](https://scientific-python.org/specs/spec-0001/) (see also [PEP 690](https://peps.python.org/pep-0690/) and the [GH repo](https://github.com/scientific-python/lazy_loader))
- [SPEC 2 — API Dispatch](https://scientific-python.org/specs/spec-0002/)
- [SPEC 3 — Accessibility](https://scientific-python.org/specs/spec-0003/)

> Q: When do you declare victory over a spec?
A: List of core project that endorse the project/specs.

[Core projects](https://scientific-python.org/specs/core-projects/)

[Steering committee](https://scientific-python.org/specs/steering-committee/)

Stefan: On the spec, there is no enforcibility. Signal to the comunity about idea that have been adopted/discussed by core projects.
It is now easier for big projects to discuss/try practices. This provides a mechanism for smaller projects to suggest ideas and have those adopted by bigger projects.

Comment from the audience: Quite similiar to the packaging PEPs (as opposed to the core python PEPs).

A: Many spec will likely not be technical, of course there are some that will be like the lazy loading.

Q: Mention about conda packaging there are [Conda Forge Enhancement Proposals](https://github.com/conda-forge/cfep), and
[Conda Enhancement Proposals](https://github.com/conda-incubator/ceps)

A: would be nice that all these coordination commitees would coordinate more

Q: it would be nice if Core Python has received feedback from the scientific Python community.
A: J:We could do better - Stefan: in the beginning of Numpy thre was more talks with Guido, and were encouraged to come up with our solutions.
Currently there is a scientific python working group at PSF, but it mostly deals with small grants only.

Q: Who is aware of the PSF funding group.
A: only a few hands up.

Should make list of funding opportunities available.
List grants, proposals and feedback, to help others write good proposals, get a sense of where money is going in the community.

Q: Dario, CZI: It would be useful if you could clarify more the scope of this proposal. Extreme use case, are you interested in supporting the long tail of library.

A: The idea of "core" project to _endorse_, as those are well followed by many developers, but we want the long tail to review and _adopt_ those proposal. Some discussion with Scikit-HEP (see their [developer pages](https://scikit-hep.org/developer)).

[Brookhaven's scientific Python cookiecutter](https://github.com/NSLS-II/scientific-python-cookiecutter)

We aim to have some central maintainer guidance at https://learn.scientific-python.org

Q: Any relationship with https://www.pyopensci.org/ ?
A: Just got funding from Sloan.
... ?
Complementary efforts.

> About pyOpenSci
>
> pyOpenSci promotes open and reproducible research through peer-review of scientific Python packages. We also build technical capacity by providing a curated repository of high-quality packages and enabling scientists to write and share their own software. We hope to foster a greater sense of community among scientific Python users so that we can help each other become better programmers and researchers.

Pangeo comment: it could be a place for collecting scientific domain stack to introduce newcomers to which kind of packages people in their domain use, it would remain some of the overwhelming effect of seeing all the scientific packages. (please do rephrase this wasn't captured very well)

**Timecheck: 7min left.**

Reminder: [Discord](https://discord.gg/vur45CbwMz)

# Puns:
- Year (day) of the linux desktop battery.