Exercise 3: Why an Old Test Restore Doesn't Prove Today's Backup Works — Possible Solution
====================================================================

WHAT THIS CHAPTER SAYS
------------------------------
Per this chapter's warn-box, "software gets upgraded, data structures
change, and backup configurations get modified over time - a test that
succeeded two years ago says nothing reliable about whether the
current backup, running against the current software, would succeed
today."

WHY A PASSED TEST IS A STATEMENT ABOUT A SPECIFIC MOMENT, NOT AN
ONGOING GUARANTEE
------------------------------
A successful test restore confirms that, at the specific time it was
run, the backup configuration in place then, combined with the
software versions in place then, produced a usable result. It says
nothing about any configuration that has since changed - which,
between the backup software, the applications being backed up, and the
underlying data structures, is essentially certain to have happened to
some degree over a long enough period.

WHY EACH KIND OF CHANGE CAN INDEPENDENTLY BREAK WHAT ONCE WORKED
------------------------------
A software upgrade can change the format a restore expects. A data
structure change can mean the backup no longer captures everything the
current version of an application actually needs (echoing this
chapter's own missing-configuration-file example, which could easily
have been introduced by a later change rather than existing from day
one). A modified backup configuration can silently narrow or shift
what's actually being captured. None of these changes require the
original test to have been wrong when it was run - they only require
time and normal, ordinary change to have passed since.

WHY "WE TESTED THIS ONCE" AND "THIS CURRENTLY WORKS" ARE DIFFERENT
CLAIMS
------------------------------
The first is a historical fact about a specific past event. The second
is a claim about the present, which requires evidence generated in the
present - or at least recently enough that none of the relevant
components have had the chance to drift since. Treating the first as
proof of the second confuses a snapshot from the past with a
continuously valid guarantee.

WHY THIS WORKS AS AN ANSWER
------------------------------
It quotes the chapter's own reasoning, explains why a test restore is
inherently a statement about one specific moment rather than an
ongoing guarantee, and gives concrete examples of how software,
data-structure, and configuration changes can each independently
invalidate a result that was genuinely correct when it was first
tested.