I’ve scanned a chunk of this conversation. I appreciate the many viewpoints and ideas.
I do think forward steps are essential, and there will undoubtedly be tradeoffs.
As someone who has spent the past decade supporting scientists with real needs for flexible, declarative packaging solutions, I face tradeoffs daily. We spend a large percentage of time supporting cutting-edge medical researchers who have real needs to use a variety of combinations of hardware, OS, and software to do research. As developers, we spend time with matrices of CI combinations to support our open source projects.
I encourage everyone to consider that “good enough” may be the best we can do. I trust that the individuals in this thread can create a “good enough” solution that is better than the “status quo”.
Thanks!
This is one example of the complexity that our users face daily:
