> If the question were, for example, what is the mass of the water contained in a 1m x 1m x 1m box, then the answer is obvious which system is by far the most sane one.
You're not disagreeing, the GP said
> If you're doing science, metric makes more sense because units work out. But most all US science has already adopted metric.
Your question asks for mass of a cubic meter of water, presupposing a typically scientific quantity (mass, instead of weight) and presupposing that a cubic meter (why not a cubic foot?) of water is a useful collection of water for some purpose. Fine, use metric. But imperial units as a system are made for practical everyday utility, which your example doesn't presuppose. On the other hand if we imperial users run across a situation where metric seems more useful, fortunately we can precisely convert, so it doesn't really matter.
In other places of utility (such as engineering disciplines) adopting a unit agnostic approach is the best. Some fields don't use either imperial or metric units but their own domain specific things, and software can always present units in whatever preference someone has or whatever is the most useful for the moment.