Personal Catalogue: React & Firebase — Chapter 1, Exercise 1
=====================================================================

TASK
Explain the real, load-bearing difference between the MongoDB sibling's
own Mongoose discriminators and a plain Firestore collection —
specifically, what guarantee discriminators give at write time that
Firestore does not, and what has to replace that guarantee in a
Firestore-based app.

SOLUTION

Mongoose discriminators give a real, enforced guarantee at write time:
before a document ever reaches the underlying MongoDB collection, it is
checked against the schema for its own declared item type. A document
tagged as a Book but missing a required `title` field, or carrying a
field that only a Dvd schema defines, is rejected by Mongoose itself —
the database never sees the bad write at all. That guarantee holds
regardless of which part of the codebase is doing the writing.

Firestore gives no equivalent guarantee. A Firestore collection is a
plain bucket of documents with no declared shape at all — any document,
with any fields, in any combination, can be written to any collection,
and Firestore's own database layer will accept it without complaint.
Nothing prevents a bug from writing a document tagged `itemType: "Cd"`
that actually carries a Dvd's own `runtime` and `director` fields
instead of a Cd's `artist` and `tracklist`.

Something else has to fill that gap in a Firestore-based app. The real
options are: validating shape in application code before every write
(easy to get right once, easy to forget in a second code path later);
enforcing structure through Firestore Security Rules, which can check a
document's own field names and types as part of deciding whether a
write is even allowed; or accepting the risk for a small, single-user
tool where the person writing the code is also the only person who will
ever call these write functions. Chapter 2 covers which of these this
course actually reaches for.

WHY THIS WORKS AS AN ANSWER
----------------------------
It names the specific, concrete guarantee Mongoose discriminators
provide (a rejected write before the database ever sees it) rather than
a vague "MongoDB has schemas" claim, correctly identifies that Firestore
provides no equivalent at the database layer, and lists genuine,
specific alternatives rather than asserting Firestore is simply
"unsafe."