Exercise 1: A Bronze Tier That Neither Version Recognizes — Possible Solution
====================================================================

THE TEST
------------------------------
    order = {'items': [{'price': 60, 'qty': 1}], 'express': False}
    calc(order, 'bronze')
    calculate_order_total(order, 'bronze')

'bronze' is not a key either version's own discount logic recognizes -
calc()'s if/elif chain only checks 'gold' and 'silver'; apply_tier_
discount()'s own discount_rates dict only has entries for 'gold' and
'silver'.

RESULTS
------------------------------
    bronze tier - messy: 60 | clean: 60 | match: True
    expected: no discount applied at all -> 60

Both versions correctly apply zero discount for an unrecognized tier,
returning the raw item total (60) unchanged.

WHY BOTH VERSIONS AGREE, EVEN THOUGH THEY HANDLE THE UNKNOWN CASE DIFFERENTLY
------------------------------
calc()'s own if/elif chain simply has no branch that matches 'bronze' -
if neither 'gold' nor 'silver' matches, t is left completely unmodified,
which is functionally "no discount." apply_tier_discount()'s own
discount_rates.get(customer_tier, 0) explicitly returns 0 for any key
not in the dictionary, which is also "no discount" - arrived at through
a different mechanism (an explicit default value, rather than the
absence of a matching branch), but producing the identical observable
result.

WHY THIS IS A GENUINELY USEFUL CONFIRMATION, NOT JUST A REPEAT TEST
------------------------------
This chapter's own original four test cases only exercised tiers both
versions explicitly handle (gold, silver, and no tier at all - None).
This exercise specifically tests an UNRECOGNIZED tier value, confirming
both implementations degrade the same way when given input neither was
designed to handle by name - a genuinely different code path in each
implementation (missing if/elif match vs. a dict's own default), yet
still producing matching output.

WHY THIS WORKS AS AN ANSWER
------------------------------
A tier value deliberately outside either version's own known set is
used, both results are verified directly rather than assumed to agree,
and the explanation traces the different internal mechanism each
version uses to reach the same "no discount" outcome.