# *July* 2020

[](https://hackmd.io/MNcskNfZSIOoLb50GqVIGg)

###### tags: `2020 dev call`
Call co-ordinates: Mondays @ 15:00 Eastern daylight time (UTC -04:00) https://zoom.us/j/384435716

Previous notes: [2020 Meeting Agenda](https://hackmd.io/zljR-pZrQ0O5J_j4NZ-9yw)
Archive: [matplotlib/ProjectManagement](https://github.com/matplotlib/ProjectManagement)

#### [Needs discussion at some point](https://hackmd.io/uzWviu8zSUChq3XhI2FqPg)

# 04 August 2020

Moved to [August](https://hackmd.io/P3Hrfoy9RSKRA4-mXMKW2A) notes...

# 27 July 2020

## Agenda

### Follow ups
- [x] finalize names / responsibilties of [named roles](https://github.com/matplotlib/governance/pull/11/files). @tacaswell https://discourse.matplotlib.org/t/add-named-roles-to-governance/21424

### New Business:
- [x] 3.3.1 release
- [x] pydata global sprint https://discourse.matplotlib.org/t/pydata-global-sprint/21426
- [x] we now have access to a Apple Silicon mac mini
- [x] https://github.com/matplotlib/governance/pull/12
- [ ] more smaller better curated galleries?
- [ ] references?
- [ ] pyplot types? https://github.com/matplotlib/matplotlib/projects/10
### PRs/Issues:
- `set_ticks()` updates the limits https://github.com/matplotlib/matplotlib/issues/17136
- Should it be possible to suppress that with a new parameter `set_ticks(... , update_limits=False)`
- pro: There might be cases, in which you don't want to update
- con: This increases the API footprint and can be worked around by setting the limits after `set_ticks()`

## Notes

### Governance stuff
- @tacaswell tries to advocate for a progressive migration to a committee lead governance

### Named roles
* Release Manager
- a role for a single release stream, may be short or longer if LTS
* API Consistency Leader
- Ensuring we don't add extra spellings of the same thing
- Path forward for deprecations
* Reference Documentation Leader
* Narrative Documentation Leader
* Secretary
- Make sure meetings happen, that agenda is filled, and followed
* Community Manager

Empty line to fix formatting; fix later...

* @tacaswell generally wants to determine a specific level of commitment to the project, instead of asking too much of volunteers
* Term limits of a year allow a specific cut-off

#### Current status:
* Tabling change from project leader to community (committee) led for now
* No-one has objected to list of named roles
* No-one has objected to description of named roles
* There may be some tweaks to the description of the named roles
* Leave for another week

### 3.3.1 release
* A few issues/PRs still open
* May want someone to look at [#18037](https://github.com/matplotlib/matplotlib/issues/18037)
* Release should be possible in a day or two

### PyData global sprint
* PyData is merging all their conferences into one online conference, Nov 11-15
* Organize a ~20-hour sprint with a range of time zones?
* Ask some European people to cover that range

### MacMini
* We have the Apple Silicon Development Kit available on our MacStadium system
* Email @tacaswell for access if needed
* do any of our mac devs know how to get x-forwarding _out_ of a mac mini?

### Governance issue 12 - Twitter Persona
* Merge as is, as a description of what's currently going on?

### PR/Issues

#### [`set_ticks()` updates the limits](https://github.com/matplotlib/matplotlib/issues/17136)
* should be consistent behaviour, or use a flag, not context-dependent (i.e., depending on whether a limit were set explicitly)
* general consensus to leave it as is

#### [Partial handling of `rcParams['axes.unicode_minus']` by formatters](https://github.com/matplotlib/matplotlib/issues/17967)
* defer to next week

# 20 July 2020
## Agenda
### Follow ups
- [ ] ~~finalize names / responsibilties of [named roles](https://github.com/matplotlib/governance/pull/7/files).~~ (Next week)
#### Small issue label cleanup:
- [x] replace `default changes` with `API changes`
- [x] remove `not a blocker`
- [x] `enhancement` vs `new feature` usage?

#### PRs:
- [x] [Build lognorm/symlognorm from corresponding scales](https://github.com/matplotlib/matplotlib/pull/16457): #16457
- [x] MEP 22 and MEP 27
- stalled - did we really want these changes? If so, someone needs to champion and herd reviewers.
- [ ] [Color blind color cycle](https://github.com/matplotlib/matplotlib/issues/9460) when will we reconsider the defaults?
- [ ] colormap interchange format and metadata?
## Notes

### MEP 22, 27,

- lots of work sitting there, we should resolve it
#### MEP22
- discussion https://github.com/matplotlib/matplotlib/issues/8712
- most of work is merged, but not the default toolbar path.
- goal is to make the tools in the toolbar reusable in other contexts
- some design decisions need to be cleaned up
- some ability to override specific backend tools
- class spec vs string spec overwritability.
- clip board included
- Move to get into Matplotlib 3.4.
- almost all groundwork has probably been done...
- rcParam switching so little cost to making default.
- [ ] ACTION: make new default

#### MEP27
- trying to do the same thing to the rest of backends / GUI / canvas parts
- unifying the different backends to make it easier to extend them all in a uniform way.
- Should be deferred to after 3.4 / wishlist

### label clean up

- remove 'default changes' -> API changes is good enough
- remove "not a blocker"
- replace "enhancement" with "new feature"
- new feature makes new API
- performance enhancement & maintenance clean up code

### Colors and defaults
- major versions:
- MPL 4.0 major style changes? Next summer?
- User study for CZI?
- numpy and pandas doing it
- syle file repository?
- configure own style sheets
- https://github.com/dhaitz/matplotlib-style-configurator
- https://github.com/dhaitz/matplotlib-style-voting
- needs maintenance
- seeing vs distribution: `pip install cool.style`
- entrypoints/register same as colormap
- [ ] ACTION: document how to do this via modules: https://github.com/matplotlib/matplotlib/issues/14943
- can only be activated after import: can't be in `rcParams`
- clickable link on gallery?
- example already for colormaps? https://github.com/anntzer/matplotlib-cmap-template

### PRs

- #16457 was merged...

# 13 July 2020

## Agenda
- [x] 3.3.0 status / review
- [x] GSOD evaluation - we have to tell google by July 31
- [x] CZI round 3 (round 2 for us) proposal due at end of the month
- [x] review proposed named roles, open call to (self) nominate?
- [x] update on grant work
- [x] PRs
- [x] [NaNs in Paths](https://github.com/matplotlib/matplotlib/pull/17885)
- [x] re-use gridspecs for `subplot` and `subplot2grid` https://github.com/matplotlib/matplotlib/pull/17347
- [x] Linking in SVG inheritance diagrams is broken (https://github.com/matplotlib/matplotlib/issues/16856#issuecomment-645681893) Should we revert to PNG diagrams for 3.3?

## Notes

### Season of Docs:
- due 31 July
- Google gets reccs from us, google gets back w/ how many projects we get

### CZI round 3: grant for second year
- Elliott and Hannah helping w/ writing
- Discussion of metrics and progress for CZI

### 3.3
- gtk fix
- nan-PR
- docs
- Revert SVG inheritance diagrams to PNG due to link bug

### docker
- made dev-ops progress at the sprints, have a dev-ready container, can build docs

### named roles

https://github.com/matplotlib/governance/pull/7/files

- release manager(s) - Elliott?
- narrative documentation manager
- reference documentation manager - Bruno?
- community manager - Hannah (if and only if nobody else volunteers)
- developer relations - discourse/twitter question traffic control/etc
- social media/communication
- api consistency leader - Tim?
- secretary
- [ ] *Action*: Tom will write description and send out call for volunteers

# 6 July 2020
## Agenda
- [ ] Apps due 9th, last pre say: https://discourse.matplotlib.org/tag/gsod
- [ ] Is it ok to not mention the dev mailing list as a possible way for feature requests? (https://github.com/matplotlib/matplotlib/pull/17842)
- [ ] [Constrained Layout rewrite (#17494)](https://github.com/matplotlib/matplotlib/pull/17494)
- [ ] Spelling of proposed "[:type:](https://discourse.matplotlib.org/t/improving-feature-discoverability-by-centralizing-documentation-of-implicit-types/21263)" Sphinx role
- [ ] Scipy Update?

## Notes

### GSOD

- Apps due Thursday
- A few proposals

### How to contribute:
- dev mailing list for features?
- should go to github?
- process for new feature to be discussed
- Yes, just guide people to github

### New `constrained_layout`
- some discussion
- would be good to get in soon so it can work in master over the next few months.

### Spelling `:type:`

- Issue:
- things like Line2D's [`linestyle` arg](https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.lines.Line2D.html) and LineCollection's [identical arg](https://matplotlib.org/3.2.1/api/collections_api.html#matplotlib.collections.LineCollection) are redundant bits of docs, want to link to something like: https://matplotlib.org/3.2.1/gallery/lines_bars_and_markers/line_demo_dash_control.html instead of a random method's docs.
- Also things like Bbox/extents/etc are not always documented everywhere they are used: https://twitter.com/ecpensiveLife/status/1278994553111011328
- color does this
- issue with plain-text version of docs versus sphinx-directive
- glossary module?
- would need to teach
- include a hint?
- variable docstring interpolations mappings?
- concerns about injecting too much ceremony / sphinx details into the docstrings
- consensus that "glossary" approach seems like it's a good approach initiallly
- since the important part is consolidating the docs themselves, we can discuss more about exact "spelling" once we have this
- Jody: should we instead make sure that the individual class that *uses* these individual "concepts" gets a glossary?
- The "glossary" appears to have grown into the idea that we should actually have dedicated Class's for these individual concepts, so that we can centralize not just the documentation (without the need for Sphinx roles), but also have the class own the validator and rcparam serializer.
- For GSOD, I'll propose primarily the actual centralization of the documentation as a first step, then the "glossary" style with some comments about alternate approaches.

.. _supported linestyles

   asdf 
   dsa

### Scipy Update
* https://github.com/QuLogic/scipy2020-mpl-update/
* add link to new whats new
* add string formatter
* https://github.com/matplotlib/cheatsheets/