Allow DataArray objects without named dimensions?

[shoyer]

At PyData SV, @mrocklin suggested that by default, array broadcasting should fall back on numpy's shape based broadcasting. This would also simplify directly constructing DataArray objects (#115).

The trick will be to make this work with xray's internals, which currently assume that dimensions are always named by strings.

[shoyer]

As I think more about this one, I'm questioning the value of allowing for unnamed dimensions, as opposed to simply defaulting to dimension names like "dim_0", "dim_1", etc.

In particular, unnamed dimensions are not very useful on object which use indices for alignment. To do sensible numpy like broadcasting, you need to be able to insert size 1 dimensions. But since we always have tick labels, we can't automatically expand those size 1 dimensions to the size of the dimension in the other array.

I suppose we could also make tick labels optional, which would make reshape operations or np.newaxis sensible. But I think I would rather still default to np.arange(size) like pandas.

[shoyer]

Closing this as "won't fix"