By request: Interesting "mistakes" in computer organization. Most of these seemed to make sense at the time. Pentium FP stack segment registers Other HW VAX: calls, savem, lack of use bit early Alpha: no atomic ld/st byte KSR1: rings, locks PC/RT: FP inconsistency TLB shootdown -- handshake for use/dirty writeback delayed branches register windows lack of atomic RMW instructions, e.g. on NS32032 visible pipeline state Itanium, ia432 C = v. == Unix gets fork