Attribute or method?

[pcchen]

I wonder if we have a convention to use attribute or method (when both are possible). For example, in numpy one uses array.shape, while for UniTensor one uses UniTensor.shape(). Other examples in Cytnx, we use UniTensor.at([2]).value and we use UniTensor.[2].item(). I wonder is there a way to know if () needed in the end.