What I find interesting was that ERMA was totally obsolete when delivered. By the end of 1959 IBM had delivered the 1401 and the 7090 both transistorized. The 1401 was the mainstay of the business world until the advent of the 360 in 1965. The 1401 and 7090 were q dynamite combination where the 1401 served as an offline peripheral controller for the 7090. The 7094 was my first computer. Baseball was my first computer game, played on the console switches and displayed on the accumulator and index register displays. (My first video game was Space Wars on a PDP 1.)
The hardware that went into ERMA may have been obsolete, but that is only the foundation of the technology. The project was years in the making. I can imagine that trying to sell adoption of emerging transitory technology into a massive financial system would have seemed an unwarranted risk.
Magnetic ink doesn't care who reads it and magnetic ink readers don't care if they're driven by vacuum tubes, transistors, microprocessors, or (we hope) photonic integrated circuits. The vision that went into this project was phenomenal. Correct me if I'm wrong, but isn't this the foundation of our modern electronic banking?
Another point on obsolescence: consider microprocessors in satellites and probes. I understand that well proven circuitry is required, which by nature is obsolete because it trails newer system in features and performance.
"Traditionally, space-qualified single board computers (SBCs) have trailed the cutting-edge commercial and military products by factors of performance often times in excess of 10-100x."
So the technology was obsolete for "new technology" but it still was suitable for the purpose it was intended, whether finance or space exploration, both areas we don't want to make mistakes.
In 1959 the whole computer industry was on the cusp of the conversion from vacuum tubes to transistors, but it was not like jumping off the cliff. The vacuum tube predecessor to the 7090 was the 709. Vacuum tubes were unreliable because they burned out frequently. The transistor machines were a huge step forward in reliability and the 7090 was at least 20 times faster. Once the transistor machines came out the vacuum tube ones were dead. I imagine BofA was faced with the decision of a year's delay to get the new systems and rewrite the software or go ahead with what they had.
Those were the days where every new machine required a rewrite of all your software. OTOH there wasn't very much of it. ERMA had only 4000 bytes of memory and the 7090 had 32K 36 bit words.
Processors for satellites are entirely different animals. They have to be radiation hardened. It's not a matter of waiting for generation x of the x86 to prove itself. Commercial processors flat out wont work in high radiation environments.
Strikingly like real war. The only way to win is to not play.
At any rate, it seems like computer professionals have more nostalgia than those in other fields. We can still look back in recent history to people who made, you know, real advances, rather than fart games for the newest telephone.
