astropy-APEs icon indicating copy to clipboard operation
astropy-APEs copied to clipboard

Published 20 hours ago • verified publisher icon astropy

APE 18: Replace NEP 29 with SPEC 0

Open pllim opened this issue 1 year ago • 13 comments
trafficstars

As proposed by @bnavigator here:

  • https://github.com/astropy/astropy/pull/16376

For reference:

  • https://numpy.org/neps/nep-0029-deprecation_policy.html (I don't see any deprecation notice here though)
  • https://scientific-python.org/specs/spec-0000/

pllim avatar May 03 '24 14:05 pllim

  • (I don't see any deprecation notice here though)

It's the "Note" box right at the top.

This NEP is superseded by the scientific python ecosystem coordination guideline SPEC 0 — Minimum Supported Versions.

bnavigator avatar May 03 '24 14:05 bnavigator

Oh, haha... I was blind...

pllim avatar May 03 '24 14:05 pllim

APE 18 refers to NEP 29 which in turn refers to SPEC 0. So APE 18 is referring to SPEC 0 already, it's just doing it indirectly. I don't see why we shouldn't update APE 18 to refer to SPEC 0 directly.

eerovaher avatar May 06 '24 15:05 eerovaher

One less link to click in the future.

pllim avatar May 06 '24 15:05 pllim

Should we go further and see if we can "endorse" SPEC 0 and get Astropy on its schedule?

Endorsed by: ipython, matplotlib, networkx, numpy, scikit-image, scipy, xarray, zarr

nstarman avatar Jun 29 '24 21:06 nstarman

I am not sure if Scientific Python is even looking for our endorsement. We are not one of the "core libraries" in their ecosystem.

pllim avatar Jul 01 '24 14:07 pllim

Well it could be argued that astropy is on par withzarr in terms of PyPI download stats: astropy gets ~40k downloads a day, zarr gets ~25k

neutrinoceros avatar Jul 01 '24 18:07 neutrinoceros

(but zarr probably has much more room to grow)

neutrinoceros avatar Jul 01 '24 18:07 neutrinoceros

I don't want to start that fire. Haha. astropy is very domain specific, unlike zarr. 😅

pllim avatar Jul 01 '24 18:07 pllim

So it seems like the consensus is to update the APE but also I have not heard from the APE 18 authors. @Cadair and @astrofrog , what do you think?

pllim avatar Jul 01 '24 18:07 pllim

Good with me.

Cadair avatar Jul 02 '24 09:07 Cadair

@bnavigator , are you still interested to update the link? But this time, it would be in https://github.com/astropy/astropy-APEs/blob/main/APE18.rst file instead in this repo. Thanks!

pllim avatar Jul 02 '24 16:07 pllim

I am not sure if updating an accepted APE is the correct way. I originally wanted to update to whatsnew of Astropy 6.1 directly (https://github.com/astropy/astropy/pull/16376) and still think this would be more appropriate.

I tend to side with https://github.com/astropy/astropy-APEs/issues/101#issuecomment-2096270409 for the APE and would move to SPEC-0 for all future references (including whatsnews/changelogs)

APE 18 refers to NEP 29 which in turn refers to SPEC 0. So APE 18 is referring to SPEC 0 already, it's just doing it indirectly. I don't see why we shouldn't update APE 18 to refer to SPEC 0 directly.

bnavigator avatar Jul 12 '24 16:07 bnavigator