Discussion
In 2020, I reran computational experiments made in 2006 for my PhD, and packaged the whole analysis-to-paper-production pipeline with Guix:
https://gitlab.inria.fr/lcourtes-phd/edcc-2006-redone
https://dx.doi.org/10.5281/zenodo.3886739
A fellow #Guix hacker reran this Guixified pipeline days ago and it still works as intended.
@civodul Sometimes one wants to reproduce the same results using the same code and the same data. Or still the same code but with different data, so it's important to validate that the old code works as a first step.
But another dimension is to be able to run the old code on top of a modernized stack.
The Go Project has made innovative strong promises on that matter.
https://go.dev/doc/go1compat
#golang has proven its stance is helpful to maintain infrastructure code as well as business code...
@civodul ... over long periods of time (well, go 1 is only 13 years old).
As reproducibility is so important, API compatibility guarantees should be considered too as base principles, so research communities should be driven toward development environments/ecosystems with those qualities.
A space for Bonfire maintainers and contributors to communicate
