Meta-issue: Moving discussion location and reframing focus

[jthielen]

In the past year after the initial duck array coordination meeting, there has been little engagement on this GitHub repo, and other than internal work in xarray and xarray companion packages (which, certainly, has been a lot), there has been little progress in nested array compatibility and interaction in the ecosystem. For those still watching this repo, would a more productive avenue to move these discussions forward be to raise these topics on https://discuss.scientific-python.org/, which has seen quite a bit of engagement recently, and close out this repo?

Also, with the progress on the Array API standards, I believe that the issues here that remain in need of resolution are centered around arrays that wrap other arrays (e.g., xarray, pint, dask) and not duck arrays in general. While I don't have any immediate ideas for a catchy name to use like "duck array" was (wrapping arrays? container arrays? meta arrays?), perhaps we could gain more traction by narrowing the focus?

[TomNicholas]

Agree that scientific-python.org now seems to be a better place to have some of these discussions.

…

On Wed, Sep 21, 2022, 5:08 PM jthielen ***@***.***> wrote: In the past year after the initial duck array coordination meeting, there has been little engagement on this GitHub repo, and other than internal work in xarray and xarray companion packages (which, certainly, has been a lot), there has been little progress in nested array compatibility and interaction in the ecosystem. For those still watching this repo, would a more productive avenue to move these discussions forward be to raise these topics on https://discuss.scientific-python.org/, which has seen quite a bit of engagement recently, and close out this repo? Also, with the progress on the Array API standards <https://data-apis.org/array-api/latest/>, I believe that the issues here that remain in need of resolution are centered around arrays that wrap other arrays (e.g., xarray, pint, dask) and not duck arrays in general. While I don't have any immediate ideas for a catchy name to use like "duck array" was (wrapping arrays? container arrays?), perhaps we could gain more traction by narrowing the focus? — Reply to this email directly, view it on GitHub <#7>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AISNPI6CGBN2QABQCZ6TAD3V7N2NZANCNFSM6AAAAAAQSOL2HQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[TomNicholas]

See also the SPECS - issues in this repo would all make for good SPECs

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

[rgommers]

This scope narrowing seems logical to me, and may be helpful.

While I don't have any immediate ideas for a catchy name to use like "duck array" was (wrapping arrays? container arrays? meta arrays?), perhaps we could gain more traction by narrowing the focus?

Meta arrays is the term I've heard most often. Container arrays works too; wrapping arrays seems less nice to me.

[SimonHeybrock]

I am trying to start such a discussion here: https://discuss.scientific-python.org/t/multiple-duck-array-layers-composability-implementations-and-usability-concerns/552

[jthielen]

With https://discuss.scientific-python.org/t/creating-community-standards-for-meta-arrays-arrays-that-wrap-other-arrays/563 and @SimonHeybrock's https://discuss.scientific-python.org/t/multiple-duck-array-layers-composability-implementations-and-usability-concerns/552, I think we can consider this repo archivable? I tried to avoid directly referencing any of the contents here (and instead reproduce them as needed).