# PlasmaPy community meeting agenda and minutes - Tuesday 2020 October 27 at 18:00 UT

[](https://hackmd.io/ypUWLrv0RdixRNGMMMR6iA)

### Video Conference Information
* [Jitsi video conference link](https://meet.jit.si/plasmapy) with [call-in information](https://meet.jit.si/static/dialInInfo.html?room=plasmapy)
* Instant messaging: [Matrix](https://riot.im/app/#/room/#plasmapy:openastronomy.org) and [Gitter](https://gitter.im/PlasmaPy/Lobby)
* [PlasmaPy google directory](https://drive.google.com/drive/folders/0ByPG8nie6fTPMEIxTlZLZjdjYms?usp=sharing) ([minutes/agendas](https://drive.google.com/drive/folders/0ByPG8nie6fTPV1FQUEkzMTgtRTg?usp=sharing), [documents](https://drive.google.com/drive/folders/0ByPG8nie6fTPYzk2TEhTa1N6R0U?usp=sharing))
* [PlasmaPy on GitHub](https://github.com/PlasmaPy/plasmapy) ([pull requests](https://github.com/PlasmaPy/plasmapy/pulls), [issues](https://github.com/PlasmaPy/plasmapy/issues))
* [PlasmaPy Enhancement Proposals on GitHub](https://github.com/PlasmaPy/PlasmaPy-PLEPs)
* [PlasmaPy Google Calendar](https://calendar.google.com/calendar?cid=bzVsb3ZkcW0zaWxsam00ZTlrMDd2cmw5bWdAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ)

## Agenda (please feel free to edit or add items)

1. Introductions
2. Preparing posters & talks for APS DPP meeting in two weeks
1. Mini-conference!
4. Time of community meeting during/after time zone changes
5. Recounting recent discussions with people interested in PlasmaPy
6. Issues
1. ...
2. ...
7. Pull requests in progress
1. [Fields & grids (#909)](https://github.com/PlasmaPy/PlasmaPy/pull/909)
2. [Shocks (#864)](https://github.com/PlasmaPy/PlasmaPy/pull/864)
3. [Ionization state refactor & `particles` clean-up (#796)](https://github.com/PlasmaPy/PlasmaPy/pull/796) — low priority

## Minutes

* Erik made suggestions that shall be considered harmful
* Python has a [package for goto statements](https://pypi.org/project/goto-statement/)
* Peter's proton radiography
* Fields & grids
* Peter Heuer's willing to help with ParticleTracker
* Next meeting probably on Nov 16
* Not much is going to happen until
* Will skip community meeting on Nov 9 because of APS DPP meeting
* Cython vs. Numba
* Perhaps start with using Numba since it's mostly adding a decorator, and then if we run into any problems, reconsider it
* Numba probably works on any platform likely to be used...seemingly even a Raspberry Pi
* Does Numba have a performance profiler?
* Probably. One of their devs is making progress.
* Numba is seeing more development than Cython.
* Numba takes advantage of advances in LLVM and JIT, while Cython
* **We will proceed with prototyping/implementing `numba` and only change course if it becomes absolutely apparent to do so.**
* https://github.com/poliastro/poliastro does use numba - see if they have any difficulties
* Theano - old Python machine learning, hard to debug, like Dask. pymc3 took over maintenance of theano. wanted to switch over to tensorflow, but tensorflow also has difficulties. Went with Jax instead - speedups, GPU, take advantage of all the progress in JIT.
* Moral of this story: numba has the power of current development in JIT behind it, while cython doesn't really get more powerful with the years
* Jean-Luc Picard has a strong claim on being The Cutest of Borg
* [Dominik] Put affiliation on the slides rather on the conference website; also on zenodo
* Presentation prep strategy: put in Gdrive folder using templates that Erik made, for easier collab
* How does commenting on Gdrive work for .ppt/.pdf?
* US is going through daylight savings, need to update www.plasmapy.org is new UTC time.
* Fit functions PR
* Mostly done
* Nick did a fantastic job at code review
* Need to parametrize tests
* May need to discuss breaking up tests into smaller units
* Breaking up tests into smaller units is usually a good thing, but doing this too much can reduce readability. If a test shouldn't be run if a prior test failed, then putting them in logical blocks is fine. It's a balance.
* `@modify_docstring` — prepend and postpend append lines to docstring.
* `AbstractFitFunction` subclasses will have parts of the docstring be the same