What should the behavour of map and collect be?

[oxinabox]

it is a more contentious feature ...

If you believe map & collected generators ought not to simply propagate names along, then perhaps open a discussion issue?

Originally posted by @mcabbott in #73 (comment)

[mcabbott]

OK, my take is that all of these ought to propagate names:

A = [f(x) for x in nda]
B = map(f, nda)
C = [g(x,y) for x in nda, y in ndb]  # (names(nda)..., names(ndb)...)
D = map(g, nda, ndb)                 # unify(names(nda), names(ndb))
E = filter(h, nda)

Only some of them do now. What's the argument against any of these? Or what other cases need to be worried about before fixing them?

[oxinabox]

I am coming around to the the idea.
map(f, ::Tuple)::Tuple) afterall.

And also recently @glennmoy had code that would have benifited from this