Inconsistency between .summary() and .name()

[duncanwp]

For Cubes with no identifier other than a stash code the Cube.summary() method will fall back on printing the STASH attribute, if it's present. I would suggest this is done in the CFVariableMixin.name() method instead so that this logic is consistent used more widely. I'm happy to put together a pull request if this would be useful.

[pelson]

Sounds good to me. The relevant line is https://github.com/SciTools/iris/blob/master/lib/iris/_cube_coord_common.py#L96:

return self.standard_name or self.long_name or self.var_name or default

There is nothing CF-like about the .name() method, so let's make it a little more useful for UM data. 👍