The fact that heroic gestures and dedicated volunteers in such projects have occasionally resulted in some reasonably good, maybe even once in a while brilliant outcomes does not mean that, structurally, this is a good way to do things.
When I am crafting small things like https://github.com/glyph/datetype or https://github.com/glyph/fritter or https://github.com/glyph/dbxs or https://github.com/glyph/automat/pull/137 these are not "maintenance" of existing platform components, but (small!) attempts to do things in newer, better ways