(Note that the reason for explicit attribute access, rather than using the P5["foo"] scheme, is that we may well have an object that acts like a hash or array, so we can't coopt the array/hash access method for attributes. But I digress, though I didn't explain properly earlier)
Thought about that too and your argument is very reasonable. Does that also mean that we have no shortcuts for setting an attribute to a native item (set PObj["a"], 1) or is that just postponed? And what about named access of arguments, *if* I just done't care about speed?
leo
