> Then I don't think nix-as-config-only is a good move.
Yep, that was the only narrow point I was trying to make initially :)
> If a multitool's screwdriver is just as good as a normal screwdriver (granted they're usually not), why not prefer the multitool?
Under the assumption that the multitool's screwdriver is just as good, then yeah, that might make sense. But as you grant, that is usually not the case. And my original point is that this is one of those usual cases where it is not as good, not one of those rare cases where it is just as good. And the reason Nix is not just as good for this, is that it was not made for this, it was made for something different and bigger.
> For instance we have a lot of python in our stack so I've found a library called mergedict that gets us close enough to the composability of nix modules without adding a new language.
Things like mergedict, while being a common genre of solution to this problem, are not a good solution to it. They are the total mess I said I'm always keeping an eye out for better alternatives to.
I don't know yet if I think Pkl (or Cue or any of the rest) actually fit that bill, but your contention that there is no point looking for solutions to this problem short of Nix strikes me as a classic case of making perfect the enemy of good.
Yep, that was the only narrow point I was trying to make initially :)
> If a multitool's screwdriver is just as good as a normal screwdriver (granted they're usually not), why not prefer the multitool?
Under the assumption that the multitool's screwdriver is just as good, then yeah, that might make sense. But as you grant, that is usually not the case. And my original point is that this is one of those usual cases where it is not as good, not one of those rare cases where it is just as good. And the reason Nix is not just as good for this, is that it was not made for this, it was made for something different and bigger.
> For instance we have a lot of python in our stack so I've found a library called mergedict that gets us close enough to the composability of nix modules without adding a new language.
Things like mergedict, while being a common genre of solution to this problem, are not a good solution to it. They are the total mess I said I'm always keeping an eye out for better alternatives to.
I don't know yet if I think Pkl (or Cue or any of the rest) actually fit that bill, but your contention that there is no point looking for solutions to this problem short of Nix strikes me as a classic case of making perfect the enemy of good.