Take co-authoring contributions into consideration

[seisman]

Continue discussions in #3730 (comment).

Currently, we use the git shortlog -sne command to count author contributions. Below is a trimmed output on 2025/01/05:

  1054  Dongdong Tian <[email protected]>
   278  Yvonne Fröhlich <[email protected]>
   266  dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
   261  Wei Ji <[email protected]>
   256  Leonardo Uieda <[email protected]>
   192  Will Schlitzer <[email protected]>
   130  Michael Grund <[email protected]>
    68  Meghan Jones <[email protected]>
    33  Yao Jiayuan <[email protected]>
    30  Meghan Jones <[email protected]>
    16  Michael Grund <[email protected]>
    15  Liam Toney <[email protected]>

The GitHub contributors page (https://github.com/GenericMappingTools/pygmt/graphs/contributors) shows totally different numbers of commits, likely because GitHub also considers co-authorship. I feel these numbers (including the number of additions and deletions) make more sense.

To consider co-authorship, we can use the following command:

git shortlog -sne --group=author --group=trailer:co-authored-by

The output is:

  1396  Dongdong Tian <[email protected]>
   603  Wei Ji <[email protected]>
   398  Yvonne Fröhlich <[email protected]>
   266  dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
   258  Leonardo Uieda <[email protected]>
   256  Michael Grund <[email protected]>
   237  Will Schlitzer <[email protected]>
   129  Meghan Jones <[email protected]>
    53  Yao Jiayuan <[email protected]>
    45  Meghan Jones <[email protected]>
    22  Michael Grund <[email protected]>
    16  actions-bot <[email protected]>
    15  Liam Toney <[email protected]>
    14  Max Jones <[email protected]>
     8  Meghan Jones <[email protected]>
     7  Max Jones <[email protected]>
     7  Wei Ji <[email protected]>
     7  yohaimagen <[email protected]>
     5  Tong <[email protected]>
     4  TIAN Dongdong <[email protected]>
     4  Yvonne Fröhlich <[email protected]>
     3  Andre L. Belem <[email protected]>
     3  Dongdong Tian <[email protected]>

After merging the commits from the same author but with different git commit author names or emails, the numbers are the same as those of GitHub.

Currently, we count the contributions (using git shortlog -sne) in the following places (xref: https://github.com/GenericMappingTools/pygmt/blob/main/AUTHORSHIP.md):

• Changelog for each release
• Authorship on Zenodo archives of releases
• Scientific publications (papers)

I think in all these cases, the co-authorship contributions should be considered.

[yvonnefroehlich]

Thanks, @seisman, for writing up this issue!

I remember I recognized the difference between git shortlog -sne and the GitHub numbers already some time ago, but could not find the related comment (#2651 (comment)) directly.

I agree on considering co-authorship as it is fair to acknowledge any kind of review (bringing up or participation in discussions, providing ideas, but also guidance during working on more complicated tasks / learning to work on new tasks). The number of additions and deletions may partly help to consider the different work loads of PRs.

However, I am unsure, if it is possible to change the rules for the author order now, and if we have to contact and ask all (at least the affected) authors for their opinions or being OK with this?

In case we keep git shortlog -sne, I still would allow for some adjustments, e.g. @weiji14 is definitely before my in the author list!

[seisman]

Using git shortlog -sn --group=author --group=trailer:co-authored-by (without the -e option) and merging the counts for the same authors, the result is:

  1429	Dongdong Tian
   617	Wei Ji
   405	Yvonne Fröhlich
   285	Michael Grund
   258	Leonardo Uieda
   237	Will Schlitzer
   203	Meghan Jones
    55	Yao Jiayuan
    15	Liam Toney
     7	yohaimagen
     6	JingHuiTong
     3	Andre L. Belem
     3	Tyler Newton
     3	kmaterna
     2	Abhishek Anant
     2	Amanda Leaman
     2	Andrey Shmakov
     2	Claire
     2	Claire Klima
     2	Claudio Satriano
     2	Conor Bacon
     2	Jamie J Quinn
     2	Malte Ziebarth
     2	MichaeINeumann
     2	Nathan Loria
     2	Noor Buchi
     2	Shivani chauhan
     2	alperen-kilic
     2	carocamargo
     2	daroari
     2	obaney
     2	srijac
     1	Alicia Ngoc Diep Ha
     1	Andrés Ignacio Torres
     1	Becky Salvage
     1	Brook
     1	Cody Woodson
     1	Emily McMullan
     1	Jack Beagley
     1	Julius Busecke
     1	Kadatatlu Kishore
     1	Loïc Houpert
     1	Mark Wieczorek
     1	Megan Munzek
     1	Nathandloria
     1	Philipp Loose
     1	Sean H
     1	Soham Banerjee
     1	Tawanda
     1	Vitor Gratiere Torres
     1	mdtanker
     1	shahid-0
     1	teriyakiSauce23

Then, for the authorship of Zenodo releases, the order should be (based on the counts only):

  1429	Dongdong Tian
   617	Wei Ji
   405	Yvonne Fröhlich
   285	Michael Grund
   258	Leonardo Uieda
   237	Will Schlitzer
   203	Meghan Jones
    55	Yao Jiayuan
    15	Liam Toney
     7	yohaimagen
     6	JingHuiTong
     3	kmaterna
     3	Andre L. Belem
     3	Tyler Newton
     2	Abhishek Anant
     2	Jamie J Quinn
     2	Malte Ziebarth

For comparison, the current author order of the latest release is:

pygmt/README.md

Lines 141 to 158 in b46274b

	author = {Tian, Dongdong and
	Uieda, Leonardo and
	Leong, Wei Ji and
	Fröhlich, Yvonne and
	Schlitzer, William and
	Grund, Michael and
	Jones, Max and
	Toney, Liam and
	Yao, Jiayuan and
	Tong, Jing-Hui and
	Magen, Yohai and
	Materna, Kathryn and
	Belem, Andre and
	Newton, Tyler and
	Anant, Abhishek and
	Ziebarth, Malte and
	Quinn, Jamie and
	Wessel, Paul},

Leo was promoted for his early work on the project; Liam and JingHui were promoted for their works on the workshops/talks.

However, I am unsure, if it is possible to change the rules for the author order now, and if we have to contact and ask all (at least the affected) authors for their opinions or being OK with this?

It seems only Michael and Will are affected.

[seisman]

BTW, in the authorship guides, we already mentioned that reviewing PRs should be taken into considerations:

pygmt/AUTHORSHIP.md

Line 51 in b46274b

and team programming efforts (including reviewing PRs).