Coordination meeting 2023

[hamogu]

At the Coordination meeting 2022 we decided:

• To have the 2023 coordination meeting in spring (May?)
• A large majority favors a hybrid meeting 12 votes vs 6 votes for fully in-person
• location: Cambridge, MA, USA (MIT or CFA)

Potential SOC:

• (if in UK, which it's not this year): Stuart Mumford (maybe) @Cadair
• Pey Lian (yes) @pllim
• Tom Aldcroft (yes, LOC) @taldcroft
• Leo Singer (maybe, in particular if in Hawaii, which it's not this year) @lpsinger
• Moritz Günther (yes, if at MIT/CfA) @hamogu
• Wilfred Gee @wtgee (?)
• Matt Craig (willing, not chair) @mwcraig

Topics to revisit (add further suggestions as comments to this issue):

• typing, see Compound units are not valid type annotations astropy#14134 (comment)

[pllim]

Might also be good time to do v6.0 planning if not too late, @astropy/astropy-project-release-team ?

[pllim]

Maybe want to add roadmap again to this meeting too, @ceb8 ?

[ceb8]

Yes, definitely Roadmap should go on the agenda. With a pre-meeting solicitation of items from the community.

EDIT: See #330

[hamogu]

Notes on physical location and hybrid mode: At the 2022 coordination meeting we discussed different options for the location. Many contributors are located in North-East America, but to be inclusive to e.g. European contributors, it's not good to default to "US East Coast". After discussion and in light of the ongoing COVID thread and related restrictive travel policies for some institutions, the US North-East is one of the places with the best probability to actually get enough in-person participation to make a hybrid meeting (the preferred option) possible this time. Maybe the 2024 meeting can be in a different place again?

At the same time, the organizing committee should take steps to make a hybrid meeting work for others. Ideas are:

• Put any discussion that might lead to important decisions into a dedicated 2h block in the morning that is easy to participate in on zoom.
• More ideas needed - please collect as comment to this thread.

[pllim]

From dev telecon today, we want to add this to potential topics to discuss during the meeting. cc @nstarman and @WilliamJamieson

• APE 2 on deprecation policy -- we need to wrap it up and merge it, see Addendum to APE 2 astropy-APEs#20

[pllim]

From Scipy 2022:

• units and constants and NEP 42 astropy#13460

[pllim]

There was a hybrid BOF at Scipy 2022. I'll leave a few notes here in case they are useful for our planning. There were some that I don't think apply to our meeting:

• For each hybrid session, have two moderators: One for on-site portion, one for virtual portion.
• Don't try to make virtual experience same as on-site; that is unrealistic. Instead, take advantage of unique opportunities that hybrid model presents.
• Enable and encourage virtual "hallway tracks".
• https://www.slido.com/

[pllim]

Sprint idea from Scipy 2022:

• Adding property-based tests for Astropy using Hypothesis astropy#9017

[eteq]

At the NF Summit this year there was a session about "contributor experience lead" - the scope of this roles in other projects as somewhat variable, but it's something in a continuum between our concept of Community Engagement Coordinator and PR first-responders. Do we want to have a similar role in Astropy? And if so, how does it overlap with our existing roles?

[pllim]

Also this:

• What is Public API astropy#13910

[pllim]

Asked by @hamogu -- Do we want to select a priority theme for the next release(s)? (In the past we uses, e.g., "performance" as a theme.) Or not worth it?

Suggestions:

• Cloud-friendly
• Dask / xarray integration etc.
• Performance again

@eteq -- Some Cycle 3 funded projects for "cloud-friendly", but maybe for next release -- v6.1?

We can discuss unless CoCo decided to send another survey out and do it that way instead. cc @adrn

[pllim]

Possible future directions of the projects; high level goals we should work towards to as an organization to stay hip.

[pllim]

• Retire this and use pydata-sphinx-theme instead? astropy-sphinx-theme#29

[hamogu]

Some ideas to integrate remote participants better into the meeting. Drawn from previous meetings I;ve attended and discussion in the Astronomers 4 Plenat Earth symposium.

• want to go beyond "listen in on zoom"
• https://tfom.org/ has a good list of resources
• We have one or two OWLs, which work pretty well in my experience for small groups (<20 or so) in person participants, not as well for larger groups.
• display remote participants on big monitor, so they are more visible and "on everyones mind"
• I've played with gathertown and https://hubs.mozilla.com/ or https://sci-an.com/. The latter two are more immersive, but have a higher technical buy-in - yet this might be the crowd where it works.

[hamogu]

Results of initial poll indicate that about 50% of those interested are likely to attend remote with a a slight preference for a later in May over early April.

[pllim]

• Add preferred reference and citation to sites.json? astropy-data#131

[wtgee]

There was some very vague talk about doing the meeting in Hawai’i and I think that's when I volunteered to potentially help out on the SOC.

I'm still potentially available to help with it being hosted elsewhere but probably couldn't take on a major role at this time.

[hamogu]

What do we want out of a User Survey? #11

[hamogu]

What direction will the core package take in the next few years? This is related to the discussion of the roadmap, but more specific to just the core package.
Future sub-packages for the core repository: In the past, there has been talk about merging some coordinated packages into the core, e.g. regions. That would make is more accessible (a lot more people install astropy itself then the coordinated or affiliated packages). What is the status of that? Are there any packages we want to include in the near future? Or do we promote those packages better? Or is it all good?

[pllim]

Finance Committee and/or budget updates.

[pllim]

Re: #255 (comment)

@taldcroft -- Is roadmap even used by Finance Committee wrt how to distribute funds? If not, what needs to change?

[nstarman]

Future sub-packages for the core repository: In the past, there has been talk about merging some coordinated packages into the core, e.g. regions. That would make is more accessible (a lot more people install astropy itself then the coordinated or affiliated packages). What is the status of that? Are there any packages we want to include in the near future? Or do we promote those packages better? Or is it all good?

Something to discuss, but we could aggregate the docs for Astropy + coordinated packages using https://docs.readthedocs.io/en/stable/subprojects.html. This way the docs for coordinates packages are tightly integrated with Astropy and easily discoverable.

[pllim]

Update: I started a meeting page at https://github.com/astropy/astropy/wiki/MITMeeting2023 . Eventually, suggestions here will be narrowed down and, if selected, would appear in the Agenda section over there. When agenda is finalized, this issue will be closed. FYI.

[hamogu]

@aaryapatil @jdswinbank @keflavich @eteq : Question for the finance committee: Since the Astropy Coordination meeting is an astropy event, how do we want to handle requests for travel support for attendees? Is it easier for you to have people contact finance individually if they need support? Should be e.g. book hotel rooms through the PEX card directly? It would be great if the finance committee could discuss the process and leave instructions on https://github.com/astropy/astropy/wiki/MITMeeting2023

We will also request modest reimbursement for coffee etc., but don't have a budget request until we know the number of participants; it's likely going to be similar to the last in-person coordination meeting (there will be fewer in-person attendees than in 2019, but prices have gone up significantly since then).

[taldcroft]

Another potential topic: using GitHub Codespaces for astropy development, esp. for on-boarding new contributors.

pllim EDIT: Also see astropy/astropy#14481

[taldcroft]

If there is interest, discuss typing.

[wtgee]

Please get in touch with Jim Weiss at NumFocus. The events team at NF should be able to help with lodging, catering, etc.

Hi @kelle, just to clarify, the individuals who wanting funding should get in touch with Jim or should the SOC?

[weaverba137]

Spectroscopy, please.

[pllim]

Spectroscopy, please.

@weaverba137 , can you please clarify? What are the topics on spectroscopy that you wish to discuss?

[weaverba137]

I would like to hear reports from the spectroscopy development teams, so that I can report back to my institution, which is very interested in this topic.

But mainly, I was surprised that no one had thought to mention it until just now.

[weaverba137]

An additional spectroscopy topic: Is APE 13 still considered current? It is now 5 years old. If it is not current, is there a current replacement? I want to be able to point people from outside the Astropy community to that document or its equivalent.

[pllim]

New suggestion: Unconference topic about governance of contracts.

[aaryapatil]

@ wtgee and others wondering how the funding process would work: We will approve all travel funding, without prior approval, which is for registered participants of the meeting and is compliant with the guidance published at https://github.com/astropy/astropy-project/blob/main/finance/process/request-funding.md#prior-to-travel (but without needing to submit a travel request)

[pllim]

Tentatively, we can discussed the proposed APE 21 if it is still not accepted by the meeting:

• APE 21: APE on ending long term support releases + Updates to APE 2 to reflect new release cycle without LTS astropy-APEs#82

[kelle]

Topic: How can we better support and encourage Project member's attendance and participation in conferences like ADASS.

[wtgee]

It would be good to have a conversation about how we can support the rest of the planet that is not in the USA or Europe, in particular large markets like India, China, Japan, etc.

[pllim]

Re: #255 (comment)

@wtgee , do you have something more concrete in mind? This came up a long, long time ago, but no one knew how to get around geopolitical limitations.

[wtgee]

Re: #255 (comment)

@wtgee , do you have something more concrete in mind? This came up a long, long time ago, but no one knew how to get around geopolitical limitations.

Honestly I'm not too sure what concrete actions there could be, I just get the feeling that there are probably a lot of people (especially students) that we could support/utilize better. I guess I'm imagining a discussion but admittedly don't have a ton to offer in that conversation.

We talked about having the coordination meeting in Hawaii as a way to potentially get more people from the Asian-Pacific involved so maybe we could just consider that for 2024.

Maybe a simple thing to do would be to make some kind of poll that directly asks the question and try to get people in those countries to fill it out. 🤷

[pllim]

Re: #255 (comment)

Thanks for the ideas, @wtgee . Let me bring them back to the next CoCo tag-up and discuss there as well, since SOC and CoCo do overlap and I think CoCo meets first. 😸

[dhomeier]

We talked about having the coordination meeting in Hawaii as a way to potentially get more people from the Asian-Pacific involved so maybe we could just consider that for 2024.

I think we did discuss scheduling the time slots for the last (virtual-only) coordination meeting such as to optimise them for as many time zones as possible, but at the time had virtually no feedback from outside the main European and American zones. But we should certainly update this regularly, to make it clear that other segments of the globe will be considered as well. I believe the last remote ADASS had some slots accommodating East Asian/Australian participants, and I guess there will be more on location in South America, Africa or Asia every 3-5 years.

[dhomeier]

Other note to self, but also to others interested: more active participation in https://www.euroscipy.org/2023/program.html . That does not really get out of the North-Atlantic bubble, but I still think it would be nice to have better presence there.

[weaverba137]

Potential topic: Independently of whether we continue to use package-template or switch to the Open Astronomy Guide, is there a way to automatically update boilerplate items like dependency versions (e.g. python>3.8, tox>4, numpy<1.23), or at least a central documentation site that has "this is the set of packages and versions that are currently used to test Astropy"?

[pllim]

Re: #255 (comment)

@weaverba137 , there is currently no auto-update available. I don't think this warrants a coordination meeting slot unless you have some specific ideas you want to hack on?

[weaverba137]

I was thinking about it in terms of affiliated packages, and I'm pretty sure that is a topic. If you'd prefer, I could bring it up under that topic.

[pllim]

Re: #255 (comment)

@weaverba137 , the entry about affiliated package I can find thus far is about "how to become affiliated" and not really about template. Dependency versions is so specific that it cannot be "global discussion" and might not even have enough traction to get a "break-out discussion" spot. You asked "is there a way" and I answered "not currently". What do you really wish to accomplish by bringing this to the coordination meeting? Are you able to rephrase this topic? Thanks!

[weaverba137]

How about this: Is there a way to know what upstream dependencies Astropy currently supports (equivalently: is being tested against), independently of tracking down the testing configuration files and reading those? If that did exist, one could conceive of packaging it in e.g. JSON or YAML format, then dropping those versions into templates of e.g. setup.cfg, tox.ini, .readthedocs.yml, ci_workflows.yml, etc.

And that's maybe an issue to open on package-template, etc., rather than a meeting topic.

[pllim]

Re: #255 (comment)

@weaverba137 , I opened issue at OpenAstronomy/packaging-guide#35

[pllim]

Any extra stuff, we can play by ear next week. Thanks, all!

https://github.com/astropy/astropy/wiki/MITMeeting2023