[QGIS-Developer] QGIS repository management

[Bourdon, Jean-François (DIF)]

[Intranet logo]

I'm far from a regular contributor to QGIS and thus not affected by this issue, but I see this review process the same way I see how peer-reviewed research papers get published (but a bit less stringent). In research, your manuscript is sent to 3 or 4 reviewers that will give comments, then you make corrections and resend the manuscript for a final review before getting published. This workflow is valid for every researcher that want to publish in a peer-reviewed journal, even for those that are also reviewer. It can be slow and frustrating at times, but having an other set of eyes on your work usually result in a better overall paper.

Software development is different than peer-reviewed research publication but getting inspiration by this centuries old process is certainly appropriate. Going back to QGIS, I see the value in having the PR of a reviewer reviewed by an other reviewer in this second fresh set of eyes that can improve on an already fine job. Every reviewer can be trusted for their expertise and judgement... but needing someone else to review your code shouldn't be seen as distrustful.

As for the conflict of interest if the reviewer works for the same company, in the academia world, the reviewer would be in conflict of interest (real or not) and would need to withdraw (the equivalent of a company being the same department of a research institute or a close collaborator). Considering that in academia this constrain has a lot to do with avoiding an idea being pushed without criticism, it might be overkill in software development where the ideas/orientations are not decided on a PR basis, but more at the PSC and with the general consensus between the devs (that's how I see it, I might be wrong).



Jean-François Bourdon, ing.f.
Analyste en télédétection
Direction des inventaires forestiers
Ministère des Ressources naturelles et des Forêts
5700, 4e Avenue Ouest, local A-108
Québec (Québec) G1H 6R1
Téléphone : 418 627-8669, poste 704304
jean-francois.bourdon at mrnf.gouv.qc.ca
mrnf.gouv.qc.ca

-----Message d'origine-----
De : QGIS-Developer <qgis-developer-bounces at lists.osgeo.org> De la part de Sandro Santilli via QGIS-Developer
Envoyé : 16 octobre 2023 10:57
À : Nyall Dawson <nyall.dawson at gmail.com>; qgis-developer at lists.osgeo.org
Objet : Re: [QGIS-Developer] QGIS repository management

On Mon, Oct 16, 2023 at 04:41:42PM +0200, Sandro Santilli via QGIS-Developer wrote:
> On Mon, Oct 16, 2023 at 09:59:35AM +1300, Nyall Dawson via QGIS-Developer wrote:
>
> > If you flip the situation, you'll see that yes, you do have trust!
> >
> > - a complete stranger CANNOT approve their own changes
> > - a complete stranger CANNOT approve other stranger's changes
> > - a complete stranger CANNOT approve an approved member's changes
> >
> > vs
> >
> > - an approved member CANNOT approve their own changes
> > - an approved member CAN approve a complete stranger's changes
> > - an approved member CAN approve a another approved member's changes
>
> That's partial trust. I'm trusted to be able to judge someone else's
> work but not to judge my own work !
>
> Beside the same-company reviews things, consider I could always ask a
> friend to file PRs for the sole purpose of me being able to merge them
> in (I cannot merge my own...). The policy is flawed to me.

My suggestion here is that having given someone "write privileges" status means the community trusts that individual to know the rules and be able to apply them, in a way that does not striclty requires someone else to guard after his/her work.

This isn't to say that reviews aren't important and that all of use should always aim at having reviews, but that the judgement abilities of the "writers" should be trusted, and if that trust is not hold anymore the "write privileges" should be revoked, following a documented procedure.

--strk;
_______________________________________________
QGIS-Developer mailing list
QGIS-Developer at lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer