Hi Ahnaf,
Thanks for reaching out on this topic. I’ve certainly spent many hours (days? weeks?) thinking about how to simplify this ecosystem for folks. In general, my preference is always to rely on existing systems as much as possible because by experience has taught me that custom/bespoke solutions are exceedingly hard to maintain. For this reason, I’d recommend leveraging existing solutions such as spack, conda, or pip, as much as possible. The effort to make a robust package manager is probably substantial because of all the edge-cases of strange system configurations. Allowing others to help with that in their development of generic tools is useful. The time invested in writing a new package manager could just as easily be spent in properly packaging a given tool for an existing package manage.
In our ecosystem (PyNE, DAGMC, etc) we have focused on conda via conda-forge, and I wish we had more resources to devote to making this work more smoothly. Some of the other active folks in the PyNE/DAGMC/OpenMC world are bigger fans of pip – and I could easily be swayed to switch our efforts to that, with a little technical digging. I know less about spack, but it is popular on larger computing platforms, so may (?) offer the advantage of being useful over a range of system sizes from individual desktops to large HPC systems.
Those are my thoughts….
Paul
P.S. Thanks for your ongoing contributions to these projects – I hope to finalize some of our build & test infrastructure soon and revisit some of your PRs.
--
-- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ --
Paul P.H. Wilson (he/him/his)
Grainger Professor of Nuclear Engineering
Chair, Department of Engineering Physics
o: 608-263-0807, c: 608-469-9615
paul....@wisc.edu
153 Engineering Research Bldg
1500 Engineering Dr, Madison, WI 53706
Zoom Meeting Room: https://uwmadison.zoom.us/j/6082630807
Zoom Phone Access: +1-929-205-6099, Access code: 6082630807
Computational Nuclear Engineering Research Group
--
---
You received this message because you are subscribed to the Google Groups "PyNE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
pyne-dev+u...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/pyne-dev/8f85fc40-3b80-4813-b070-77ca8dbbbe49n%40googlegroups.com.