Although, I might consider assembler code where every
second instruction
is a jump to be spaghetti code.
What about microcodes where <>n> bits of each instruction define the
address of the next intruction (this value is possibly modified by the
conditional branch logic). In other words every instruction includes a jump.
Actually, one microcode I deciphered was slightly worse. The next_address
field, rather than containing the bit pattern for the next instruction's
address, instead indicated which bits of the current address to change (a
'0' meant leave that bit alone, a '1' meant flip it). And part of the
this field also defined which conditon to use (if the condition logic was
enanled, fortuantely by a totally separate bit in the microode word).
[A very obscrue puzzel, what machine was I looking at?]
-tony