Sometimes you can explain a simple thing for the thousandth time, and come away with a deeper understanding yourself. It happened to me the other day with Python mutable argument default values.

This is a classic Python “gotcha”: you can provide a default value for a function argument, but it will only be evaluated once:

>>> def doubled(item, the_list=[]):
...     the_list.append(item)
...     the_list.append(item)
...     return the_list
...
>>> print(doubled(10))
[10, 10]
>>> print(doubled(99))
[10, 10, 99, 99]    # WHAT!?

I’ve seen people be surprised by this and ask about it countless times. And countless times I’ve said, “Yup, the value is only calculated once, and stored on the function.”

But recently I heard someone answer with, “it’s a value, not an expression,” which is a good succinct way to say it. And when a co-worker brought it up again the other day, I realized, it’s right in the name: people ask about “default values” not “default expressions.” Of course it’s calculated only once, it’s a default value, not a default expression. Somehow answering the question for the thousandth time made those words click into place and make a connection I hadn’t realized before.

Maybe this seems obvious to others who have been fielding this question, but to me it was a satisfying alignment of the terminology and the semantics. I’d been using the words for years, but hadn’t seen them as so right before.

This is one of the reasons I’m always interested to help new learners: even well-trodden paths can reveal new insights.