What comes from the code but has not been tried, and what is not settled.
The four SHIFT screens. The tone copy, the list, the printer and LAYOUT come from the code; they are neither built from the ROM nor checked.
The printer. The dot dump of the screen and the wave table (0x5419, 0x5516) have not been tried on a printer, nor is it known which model the sequences at 0x54B6 are for.
The disk code. The save, load and delete routines through the BDOS and the menu's DISK option come from the code; nothing has been tried with a drive.
RC741 86 5 31. The thirteen bytes at 0xBF62 give the catalogue number and, by the look of it, a date, 31 May 1986. Which date it is is not known.
Fifteen tones with a name and one without. Tone 16 is wave 1 with no envelopes and no filter: it may be the blank tone on purpose.
The padding at 0x7135. Between the last PLAY routine and the waves there are 203 bytes that are not code and nobody reads.
The cycles on a real machine. What is measured (254 and 255 per sample, with the M1 wait state) is from openMSX; it has not been measured on a real MSX.