The README had become a changelog: which session found what, which constant was
retired and by whom, which figure was overturned. That is what STATUS is for,
and it made the front page a poor read for anyone who has not been following.
Rewritten to say what the project is, what it looks like, where it stands and
how to reproduce it, present tense, no session numbers, no em dashes.
Restating the standing facts caught one of my own from this session. FINDINGS
52.5 cited 42.4's W sensitivity table -- W <= 6 fits 0/120 frames, W = 8 misses
47/120 -- as though those were clocks per BYTE. They are per WORD. FINDINGS 43
is the section that caught W being charged per word to a byte-wide port and says
in terms that 42.3's 0/120 was never physically reachable, so quoting it in byte
units re-imported the exact 2x error 43 exists to have corrected, one section
after using that same trap as a warning.
Struck in 52.5 and in ROADMAP P4, and replaced with the ladder buscost.py
already carries, which is in the right unit: 5 clk/B single address with the bus
held, 9 dual address held, 12 single address arbitrated per byte, 16..19 dual
address arbitrated per byte, the last being what the IPL ROM programs. Nothing
else in 52 depended on the struck figures -- 15_bus_occupancy.py's sweep is in
byte units and is unaffected.
The "cross-check" claim goes with it. 52.5 read the agreement between its W = 8
row and 42.4's 47/120 as two independent models landing in the same place. It is
two different units on two different containers at two different rates, and
calling it corroboration was manufacturing agreement out of the unit error.
What replaces "W <= 12 is a requirement on the player's DMAC programming" is the
sharper version: getting the DMAC to HOLD THE BUS is what separates 9 from 19,
and that is the thing P4 has to demonstrate.
Docs only; no code changed and no gate output moves.
Claude-Session: https://claude.ai/code/session_01194oWYW8DQXK1SZ2DnChW6
prosolis
ed172c2da2