ENIAC

1946 January 4

A Manual of Operation for the Automatic
Sequence Controlled Calculator

      I’ve finally completed that book Aiken instructed me to write. It’s titled “A Manual of Operation for the Automatic Sequence Controlled Calculator”, not a very snappy name, I know. It’s a five hundred paged book that’s both a history of the Mark I, and a guide to programming it. The first chapter has information about earlier calculating machines, especially those built by Pascal, Leibniz and Charles Babbage. I began with an epigraph from Babbage. Like Ada Lovelace, I understand that Charles Babbage’s analytical engine had a special quality, which I believe will distinguish the Harvard Mark I from any other computer of the time. Like Ada and Babbage’s unbuilt analytical engine, the Mark I can be programmed with new instructions, unlike the ENIAC.

ENIAC

ENIAC was built as a digital electronic computer using circuits with vacuum tubes, that solves differential equations and can perform other tasks as well. Commander Aiken had already visited ENIAC, therefore I decided to go up to Penn to see ENIAC for myself. I was unimpressed. It seems to me that the Mark I is much more practical, and Aiken agrees with me on this, because it is easily programable, whereas ENIAC could take an entire day to be reprogrammed. The end of the war brought new problems for the ENIAC as well. The machine was originally designed to perform a specific set of calculations repeatedly, but now that the need for the calculations of missile trajectory has diminished the ENIAC is being used for other tasks, such as sonic waves and weather patterns which would mean it must be reprogrammed more often. But I suppose I can see the advantages, and changes are even being made to improve the reprogramming time, and clinging to the past is not always the best option. 

Debugging

1945 March 15th

       My first compiler, the A-O series was made to translate symbolic mathematical code into machine code. It also allowed the specification of call numbers assigned to the collection programing routines stored on magnetic tape. This allowed programers to specify the call numbers of the desired routines and the computer would find them all on tape and do the calculations for you. After that, I thought why not make a language that could be understood by anyone, and not only programmers and mathematicians? 

Inside UNIVAC

   I’ve been working on developing the B-O compiler for a while now. It should be able to translate a computer language that can be used for traditional business activities like automatic billing. My staff and I were able to use B-O to make the UNIVAC I and II understand twenty statements in English, but despite this accomplishment we were told very quickly that this wouldn't be possible because computers didn't understand English. I really don’t understand the thinking behind that conclusion, we've even demonstrated that it was feasible on the UNIVAC.

    Anyway, the Harvard Mark I is officially the world's most easily programmable large computer. It can switch tasks simply by getting new instructions on punch paper, rather than having to reconfigure its hardware or cables. But even as impressive as this computer has become, it seems like before we can get the word out about her, she's become a dead duck, and now everyone is going electronic. I guess her electromechanical relays just aren't fast enough anymore.

In other news, we've been keeping the Mark II in a room with no screens on the windows. As we were inspecting the machine, because it mysteriously quit working, we found a moth, with a wingspan of four inches that was smashed in one of the computers electromechanical relays. Of course, we had a good laugh over that, then pasted it into the log book and there we have our first actually case of a bug being found. because of this incident, we now refer to removing glitches as ‘debugging the machine’.🐛

First actual case of bug being found

Mark I

1944 November 20th

     I’ve recently graduated from the Naval Reserve Midshipmen’s school at Smith College, first in my class as Lieutenant Grace Hopper. I had a few reasons to want to join the Navy. First of all, I was bored. Bored with my job as a college professor, and bored with my marriage. Second, it was sort of a tradition in my family, my great grandfather had been in the navy, and if I hadn't joined he probably would have risen from his grave and haunted me. I assumed I was going to be assigned to a cryptography and code group, since those are more in my wheelhouse, but to my surprise, I was ordered to report to Harvard University to work on the Mark I. 

Commander Howard Aiken

Now I’m working on this enormous computer called the Harvard Mark I. It’s quite impressive really, a five ton piece of machinery, all bare and open and extremely noisy. The machine that's making this huge racket is in fact a computing machine, made for gunnery and ballistic calculations by my new boss, Howard Aiken. I’ve been poring over the blueprints for days, and with some assistance from Richard Bloch I think I’ve finally got it figured out.

The Harvard Mark 1

Speaking of my partner Bloch, he and Aiken have bit of a strained relationship. I’ve been trying to tell him that Aiken is just like a computer, he’s wired a certain way, and if you’re going to work with him you just have to realize how he’s wired. Bloch just doesn’t seem to understand how to work with Aiken, which has led to him getting in trouble more often than not.

I’ve also been assigned to write a book, which may be problematic since I’ve never written a book before. One day Commander Aiken came over to my desk and said, “You’re going to write a book.” and I told him I couldn’t do that, I’d never written a book before. 

He responded, “Well, you’re in the Navy now, you are going to write one.” Which I suppose is true. 

The book is supposed to be a guide to programming the Mark I as well as a little of it’s history. Since I really have no idea what I’m doing I've been meeting every evening for feedback from Aiken and reading him the five or so pages I’d written that day. He pointed out to me, that if you stumble when you read it aloud, you'd better go fix that sentence. It works too. I’ve improved my writing with his guidance, but the book still has got a ways to go.