Exercise 2: Why the Z80 Has Roughly Double the 6502's Transistor Count — Possible Solution
====================================================================

TYING THE NUMBERS TO EACH CHIP'S OWN CONSTRAINT
------------------------------
Per cpu8bit1-2, the 6502's ~3,500 transistors are a direct consequence
of its founding constraint: a $25 price target that had to be hit from
a completely clean slate, which meant justifying every single
transistor against a tight manufacturing budget and cutting anything
not strictly necessary.

Per this chapter's own framing, the Z80's founding constraint was the
opposite kind entirely: full compatibility with the existing 8080,
which by itself already required a certain baseline of transistors
just to replicate the 8080's own functionality — PLUS a substantial
amount of genuinely new capability added on top (roughly double the
instruction count, entirely new shadow and index registers previewed
for cpu8bit1-6, more addressing modes previewed for cpu8bit1-7). None
of that new capability was optional the way it would have been for a
clean-slate design — once "double the instructions" and "new register
sets" were the goal, the transistors needed to implement them were a
direct, unavoidable cost of achieving that goal.

WHY THIS IS MORE THAN "MORE STUFF"
------------------------------
The point isn't simply that the Z80 "has more features so naturally
it uses more transistors" — it's that the two chips' transistor
counts are each the direct, traceable outcome of two DIFFERENT
optimization targets. The 6502 was optimized to minimize transistor
count itself, as the actual goal. The Z80 was optimized to maximize
compatibility-plus-new-capability, with transistor count simply
following wherever that goal required it to go — a genuinely
different relationship between "goal" and "transistor count" in each
case, not just two chips that happened to end up in different places.

WHY THIS WORKS AS AN ANSWER
------------------------------
It explains the roughly 2x difference by tracing each chip's
transistor count back to its own specific, stated founding
constraint (cost-minimization vs. compatibility-plus-addition) rather
than treating "more features costs more transistors" as a
self-evident, unexamined fact.