Yes, a rewrite is plain silliness. Never take the world by storm is more likely
to happen if that advice is followed. Moreover, I don't see any proof in actual
progress or code, for your argument. The compiler is already written in Nim.
Nim itself is an elegant language with a lot of refactoring possibilities. so
the situation of creating a messy complex, multi layer bloatware that you
imagine is, afaik, nonexistent. With incremental compilation, and other tricks,
compilation times will be faster. Then stdlib grows, ports can happen etc, but
that is aside the main compiler work. I am sure Araq et al know much better
what needs to be done. I think the addition of explicit interfaces, or fixing
the concepts feature to work better can provide a more elegant stdlib. distinct
types can make it even tidier. but Nim imho should remain simple and "lowlevel"
in a way, that is, easy to get to the actual data directly. some high level
languages make it impossible by hiding that part, so you have to use their
classes which have only partial functionality, so this layering approach can be
quite unnecessary and counterproductive.
