Plots maintenance call

[BeastyBlacksmith]

Maitainers and interested users are invited to join the maintainer call, which will happen on every ~~first Monday~~ second Tuesday of each quaterly year (such as 05.07.2021, 04.10.2021, ...) at 14:00 CET/CEST. Get in contact with me on zulip to get an invite.

See you soon!

[BeastyBlacksmith]

🔜

Meeting at 10.10.2023 14:00 CEST

• #4565

📃 Backlog:

Meeting at 04.07.2023 14:00 CEST ❌ (canceled)

Meeting at ~~11.06.2022~~ 11.10.2022 14:00 CEST

Topics

• https://github.com/JuliaPlots/Plots.jl/pull/3989
• https://github.com/JuliaPlots/Plots.jl/pull/4025 - this should be configurable with an additional attribute
• https://github.com/JuliaPlots/PlotUtils.jl/pull/92
• https://github.com/JuliaPlots/PlotThemes.jl/pull/69

Meeting at 20.04.2022 14:00 CET/CEST

Topics

• https://github.com/JuliaPlots/Plots.jl/issues/4114 - keep things as is for now

Meeting at 11.01.2022

Topics

• PlotDocs.jl is ~700MB in size
• Does making recipes fully featured require 2.0 (think @layout and friends) https://github.com/JuliaPlots/RecipesBase.jl/issues/15 https://github.com/JuliaPlots/Plots.jl/issues/4432?
• Accessor API?
• Reflect on @add_attributes (https://github.com/JuliaPlots/Plots.jl/pull/2854). Should we apply it throughout?
• Need versioned documentation

Meeting at 04.10.2021

Topics

• Series API for plots?: E.g. for a fillbetween implementation. series_plotindex kind of does it, but it's not clear whether it returns the wanted points for recipes like histogram
• might be possible to implement colorbar as plot recipe to be able to place it independently from subfigures
• Open PRs
  • https://github.com/JuliaPlots/Plots.jl/pull/3782 ✅

Next meeting at 11.01.2022 change to every second tuesday / 3 months

Meeting at 05.07.2021

Topics

• Roadmap
  • add colorbar properties: ticks and scale for major backends (GR, PyPlot, PGFPlotsX)
  • Document: rows are segments for series
  • Update supported arguments and series in the documentation
• Open PRs ✅
• https://github.com/JuliaPlots/Plots.jl/pull/2680 ✅
• the DPI issue: size shouldn't scale with dpi for vector graphics
• StatsPlots maintainer: Publicly look for new maintainer ✅

[isentropic]

Along with piled-up open PRs we should also discuss the piled up no longer valid issues https://github.com/JuliaPlots/Plots.jl/issues?q=is%3Aissue+is%3Aopen+sort%3Acomments-desc

Also, breaking some long-standing Plots defaults, like color scheme, font size, a legend by default, ... Perhaps run Julia userbase wishes poll for Plots 2.0 in the discourse. I think the plots should look good; legible; colorblind-aware as much possible by default. Currently I find that I need to apply ~5lines of customization to Plots default to it make subjectively look good to me. 95% percent of the userbase might not even know / bother how to do those default customizations, while a poweruser can easily bring everything back to default Plots 1.0 (we could even have a function legacy_styling() to bring it all back to the original design.

[isentropic]

Perhaps @fonsp could know what are the problems with the default styling of the Plots better than we do.

I personally think fontsizes should be higher, colorscheme should be more "pastel", perhaps even take out the grid, as more often that not it is not helpful (I think).

We should collect opinions of the mass, but also opinions of the big people (package devs) separately and come up with and compile a consensus. Some names that come to mind @dpsanders @fonsp ... (please put your suggestions)

Right now, I feel that Plots API is a joy to use, but the styling needs more work from.

[BeastyBlacksmith]

We should definitely make some kind of list of what needs to be done in Plots 1.x (hence the roadmap). When that is done, we can think about breaking changes, but that's a topic for a future call, I'd say.

[isentropic]

Then I suggest to discuss colorbars and fix'em up across backends and agree upon colorbar specific kws.

[fonsp]

Cool! I will try to join! The biggest problem we faced during the class with Plots.jl+GR is scaling issues: plots can randomly display very large or very small, and the dpi kwarg seems counter-intuitive (https://github.com/JuliaPlots/Plots.jl/issues/3192). It feels like there is some statefulness/system-dependence in the scaling logic. There were some patches in this area (https://github.com/JuliaPlots/Plots.jl/issues/3133 and https://github.com/fonsp/Pluto.jl/pull/1090) so this complaint might be outdated 🙊

I also have some API-related comments that come up when using Plots.jl inside Pluto.jl, like the statefulness of plot!() and default(). I have suggestions if there is interest :)

About styling: we could try to go from a "2000s look" to a "2020s look"? It might be good to create "mood board" in a Google Doc / notion.so page with screenshots from plots in the wild, where you annotate what you like or don't like about it. (I would definitely contribute!) This worked well for my design projects, but I am not sure how well it would work in a large group of people with different visions.

[fonsp]

Are you having a call today? I got a calendar invite 📅

[BeastyBlacksmith]

Yeah, I am on holidays, so I shifted the meeting to 20th of April

[isentropic]

i logged in to check but there was no meeting

[isentropic]

Could you send out a new invite here or calendar