[thaumasiotes]

> Valid use cases for `is` at all are rare.

There might not be that many of them, depending on how you count, but they're not rare in the slightest. For example, you have to use `is` in the common case where you want the default value of a function argument to be an empty list.

[zahlman]

I assume you refer to the `is None` idiom. That happens often enough, but I count it as exactly one use case, and I think it's usually poorly considered anyway. Again, you probably don't actually want the default value to be an empty list, because it doesn't make a lot of sense to mutate something that the caller isn't actually required to provide (unless the caller never provides it and you're just abusing the default-argument behaviour for some kind of cache).

Using, for example, `()` as a default argument, and cleaning up your logic to not do those mutations, is commonly simpler and more expressive. A lot of the community has the idea that a tuple should represent heterogeneous fixed-length data and a list should be homogeneous; but I consider (im)mutability to be a much more interesting property of types.

[fwip]

Could you expand on this? For example, this works just fine:

    def silly_append(item, orig=[]):
       return orig + [item]

Edit: Oh, I think you probably mean in cases where you're mutating the input list.