Semantics: axis-name vs interface-name
I think calling one of the rating axes an 'interface', as in <value interface-name="depth"> is a bit confusing.
Before (too many) people rely on it, perhaps this could be renamed?
This is only APE at the moment, or not?
Yes this is APE only. It can be renamed to axis-name. It's really only for the results XML where the metric tags are labelled with this to discern the movements caused by one axis.
Makes sense. On the other hand, the different axes are actually called 'interface' as well (i.e. you add another interface node in test_create). So perhaps this naming convention works...
It's just that normally an 'interface' is the whole thing.
Perhaps the <interface> node itself should be <axis>?
Makes sense to me. Somewhat reluctant to break this for legacy result files etc. - is it possible to make it flexible / backwards compatible for a while?