📣 Looking for tutorials and other references about containers!

In the context of #ReproducibleResearch [1], we’re writing “cards” gathering resources for more robust scientific outcomes.

You want to help?

Easy: mention one tutorial or some reference that already helped you!

• Either a practical getting started;
• Either something that helped you to get how containers can boost your research by being more reproducible.

Checkout: https://gt-env-logiciels.gricad-pages.univ-grenoble-alpes.fr/sandbox-notecards/cards/containers

Thanks

#OpenScience
1: https://www.recherche-reproductible.fr

📣 Looking for tutorials and other references about containers!

In the context of #ReproducibleResearch [1], we’re writing “cards” gathering resources for more robust scientific outcomes.

You want to help?

Easy: mention one tutorial or some reference that already helped you!

• Either a practical getting started;
• Either something that helped you to get how containers can boost your research by being more reproducible.

Checkout: https://gt-env-logiciels.gricad-pages.univ-grenoble-alpes.fr/sandbox-notecards/cards/containers

Thanks

#OpenScience
1: https://www.recherche-reproductible.fr

Personne qui interviewe : « Cela prend du temps et les chercheurs en manquent cruellement… »

🤔 Et si on réduisait la bureaucratie, euh, compétitivité, euh, de mettre des couches de labels d’excellence, euh… bah peut-être que le temps manquerait moins, non ?

Sarah Cohen-Boulakia : « Aujourd’hui, on publie trop et trop vite. Il faudrait évaluer les chercheurs davantage sur la qualité et la transparence plutôt que sur la quantité.

La peur du “décrochage de la France” en termes de publications scientifiques terrorise beaucoup de décideurs alors que la transparence nous permettrait d’avoir des résultats plus solides et de regagner la confiance du grand public en ces temps de désinformation. »

👍 😍

https://themeta.news/sarah-cohen-boulakia
#ReproducibleResearch #OpenScience

Personne qui interviewe : « Cela prend du temps et les chercheurs en manquent cruellement… »

🤔 Et si on réduisait la bureaucratie, euh, compétitivité, euh, de mettre des couches de labels d’excellence, euh… bah peut-être que le temps manquerait moins, non ?

Sarah Cohen-Boulakia : « Aujourd’hui, on publie trop et trop vite. Il faudrait évaluer les chercheurs davantage sur la qualité et la transparence plutôt que sur la quantité.

La peur du “décrochage de la France” en termes de publications scientifiques terrorise beaucoup de décideurs alors que la transparence nous permettrait d’avoir des résultats plus solides et de regagner la confiance du grand public en ces temps de désinformation. »

👍 😍

https://themeta.news/sarah-cohen-boulakia
#ReproducibleResearch #OpenScience

« Il faut expliquer le mieux possible comment un résultat a été obtenu, rendre les plus traçables possibles les conditions de son obtention.

C’est d’abord ce manque d’information qui crée la crise.

On parle de crise de la reproductibilité mais en réalité c’est la transparence qui compte et qui permettra de la résoudre. Chacun a un rôle à jouer : décideurs, éditeurs, chercheurs… »

🤔 D’ailleurs, peut-être que la crise est une crise de confiance, et dont la transparence est un moyen pour la résoudre.

https://themeta.news/sarah-cohen-boulakia
#ReproducibleResearch #OpenScience

Story: when colleagues realize #Guix rocks!

– Please help me! a colleague said.
– What’s wrong? I answered.
– My R stuff worked and now it crashes!
– Ah, so what’s changed in between?
– Nothing!
– Maybe an update?
– Indeed, apt update and upgrade, but packages were unrelated!
– Maybe not?
– I’ve double-checked:
• it still works on this other machine but crashed on this one.
• all the versions between the both machines are the sames.
Damned! I’m lost…
– Maybe it’s not the same versions?
– Look, 1.2.3-xzy and 1.2.3-xyz etc.
– Look, folder checksum are different so it’s not the same.
– …
– Please, give me the list of R packages you need.

…half hour later…

Thanks #Guix and guix-science, guix-cran and guix-bioc channels, only one R package needs to be freshly packaged.

The command:

guix time-machine -C channels.scm -- shell -m manifest.scm

and tandam the same computational environment is running on the both machines.

Nothing more.

First conclusion: APT and Ubuntu are poorly designed for #ReproducibleResearch and one needs wizard skills to avoid the bitter bite.

Second conclusion: Some people in my lab are realizing the hard way: « Aaah maybe starting a project with Guix might be slower at first but then it frees from some painful burden. »

Third conclusion: let wait the end of some pending computations. 😁

Story: when colleagues realize #Guix rocks!

– Please help me! a colleague said.
– What’s wrong? I answered.
– My R stuff worked and now it crashes!
– Ah, so what’s changed in between?
– Nothing!
– Maybe an update?
– Indeed, apt update and upgrade, but packages were unrelated!
– Maybe not?
– I’ve double-checked:
• it still works on this other machine but crashed on this one.
• all the versions between the both machines are the sames.
Damned! I’m lost…
– Maybe it’s not the same versions?
– Look, 1.2.3-xzy and 1.2.3-xyz etc.
– Look, folder checksum are different so it’s not the same.
– …
– Please, give me the list of R packages you need.

…half hour later…

Thanks #Guix and guix-science, guix-cran and guix-bioc channels, only one R package needs to be freshly packaged.

The command:

guix time-machine -C channels.scm -- shell -m manifest.scm

and tandam the same computational environment is running on the both machines.

Nothing more.

First conclusion: APT and Ubuntu are poorly designed for #ReproducibleResearch and one needs wizard skills to avoid the bitter bite.

Second conclusion: Some people in my lab are realizing the hard way: « Aaah maybe starting a project with Guix might be slower at first but then it frees from some painful burden. »

Third conclusion: let wait the end of some pending computations. 😁

« At the same time, we continue to invest enormous effort in perfecting the presentation of the final PDF like typesetting, formatting, and version-freezing, while under-investing in the earlier stages of the research journey where quality, integrity, and reuse are actually shaped. That imbalance is not a failure of intent, but a misalignment of effort driven by legacy workflows. »

🤔 Misalignment raised to power of “Publish or Perish”?

https://scholarlykitchen.sspnet.org/2026/02/05/back-to-the-article-of-the-future-an-interview-with-sami-benchekroun-and-rod-cookson/
#ReproducibleResearch #OpenScience
1/3

« At the same time, we continue to invest enormous effort in perfecting the presentation of the final PDF like typesetting, formatting, and version-freezing, while under-investing in the earlier stages of the research journey where quality, integrity, and reuse are actually shaped. That imbalance is not a failure of intent, but a misalignment of effort driven by legacy workflows. »

🤔 Misalignment raised to power of “Publish or Perish”?

https://scholarlykitchen.sspnet.org/2026/02/05/back-to-the-article-of-the-future-an-interview-with-sami-benchekroun-and-rod-cookson/
#ReproducibleResearch #OpenScience
1/3

« La leçon principale de ce projet est donc : on ne peut pas bâtir une plate-forme de calcul reproductible sur une fondation instable comme Python. »

Et la personne qui écrit cela est mentionnée dans “The Story of Python and how it took over the world” (Python: The Documentary).

Mais qui est-ce donc ? 😉

Un indice ? Aujourd’hui elle promeut (Common) Lisp. 😁

https://youtu.be/GfH4QL4VqJ0

#ReproducibleResearch

« La leçon principale de ce projet est donc : on ne peut pas bâtir une plate-forme de calcul reproductible sur une fondation instable comme Python. »

Et la personne qui écrit cela est mentionnée dans “The Story of Python and how it took over the world” (Python: The Documentary).

Mais qui est-ce donc ? 😉

Un indice ? Aujourd’hui elle promeut (Common) Lisp. 😁

https://youtu.be/GfH4QL4VqJ0

#ReproducibleResearch

The annual Guix-HPC report was published last Friday, check it out!
https://hpc.guix.info/blog/2026/02/guix-hpc-activity-report-2025

#Guix #HPC #ReproducibleResearch #OpenScience

Nice poster on using #Guix for #ReproducibleResearch 👇
https://zenodo.org/records/17435294/files/Long_lasting_bit-reproducible_research_with_GNU_Guix.pdf

Nice poster on using #Guix for #ReproducibleResearch 👇
https://zenodo.org/records/17435294/files/Long_lasting_bit-reproducible_research_with_GNU_Guix.pdf

The annual Guix-HPC report was published last Friday, check it out!
https://hpc.guix.info/blog/2026/02/guix-hpc-activity-report-2025

#Guix #HPC #ReproducibleResearch #OpenScience

@GuilloteauQ @hannes @Zimm_i48 @zacchiro @luj Amazing work.

Many of us have been repeating “Dockerfiles cannot yield reproducible images” for years, but a thorough literature review and experimental demonstration like this really hits the nail.

#ReproducibleResearch #OpenScience