Alan Kay: No, ENIAC Didn't Have a BIOS — And Wasn't Really a Stored-Program Computer

Computer scientist Alan Kay has offered a detailed answer to a lingering question in computing history: did ENIAC, one of the earliest electronic computers, have anything resembling a BIOS? His answer, posted on Quora, is a firm no.

Kay, known for his pioneering work in object-oriented programming and personal computing, wrote that he does not consider ENIAC to have been a stored-program computer at all. "There were a few tricks you could do along those lines, but not enough IMO to qualify," he said, addressing arguments that ENIAC's design included limited workarounds resembling program storage.

The question of a BIOS, Kay noted, raises a deeper issue: whether the basic input/output routines needed to start a computer must live in firmware to count as a true BIOS. "The letter of the law says yes," he wrote, but he sided with a looser interpretation that doesn't require a read-only memory chip to satisfy the concept.

To illustrate, Kay pointed to computing practice well into the 1960s, when many machines started up with no code loaded at all. Operators would manually enter a handful of instructions using toggle switches, which would then load a larger program from paper tape. This was standard procedure for early DEC computers, Kay said. The CDC 6600, introduced around 1964–1965, improved on this with a dedicated panel of switches for "dead starts," a more convenient alternative to individual toggles.

Kay also cited Wikipedia's account of the BIOS concept's formal history: the term was first officially used in 1975 for the CP/M operating system, while the version of BIOS familiar to most computer users dates to around 1981, with the launch of the IBM PC. Both of those implementations, he noted, were ROM-based — a technical detail that underscores why ENIAC, built more than three decades earlier and without any such firmware, doesn't fit the definition.