Guix 1.2.0 & reproducible research

[Ludovic Courtès]

Hello!

Some news to break the silence on this list…

On the Guix-HPC blog, we announced last week what HPC practitioners and
scientists interested in reproducible research may like from the new
1.2.0 release:

  https://hpc.guix.info/blog/2020/11/hpc-reproducible-research-in-guix-1.2.0/

A bit earlier, Konrad Hinsen wrote an article entitled “The four
possibilities of reproducible scientific computations” that in my
opinion very nicely summarizes requirements for a “reproducible
computation” and their implications:

  
https://blog.khinsen.net/posts/2020/11/20/the-four-possibilities-of-reproducible-scientific-computations/

It effectively rules out pre-built binaries (Docker images) as well as
SaaSS (Binder), but it does so in a way that demonstrates how they fail
some of the basic high-level criteria rather than entering a technical
debate.

Ludo’.