|Digi-Comp I (photo from Wikipedia)|
The first example was when I was proof reading my next title for St Martin's Press (not due out until significantly later in the year), called Ten Billion Tomorrows. The book about the relationship between science and science fiction, and I point out that when I saw 2001: A Space Odyssey in the cinema, the only computer I had ever seen before I encountered the remarkable Hal was my Digi-Comp I. This was a mechanical device with three plastic sliders, which could be programmed by adding extensions on the side of the sliders which flipped metal wires, and as a result could provide the action of different gates and reflect the outcome on 3 mechanical binary displays. Sophisticated it was not.
Examples two and three involve good old Charles Babbage. You just can't talk about mechanical computers without mentioning Babbage. He first came up in my review of James Tagg's Are the Androids Dreaming Yet, which confuses an image of the Science Machine's Difference Engine with the Analytical Engine. (The first was a hard-geared mechanical calculator, while the second, never built, was a programmable computer that would have used punched cards. Babbage built a small segment of the Difference Engine, but never got anywhere with the Analytical Engine, which probably would not have been practical given engineering tolerances.)
Then Babbage popped up again in a Guardian article about a graphic novel featuring the Analytical Engine. As Thony Christie points out in a blog post, the article wildly overstated the contribution of Ada King* to the project saying that 'Ada Lovelace and Charles Babbage designed a computer' and 'for which Lovelace wrote the programs.' In fact King had nothing to do with the design, she translated a paper on the concept from the Italian and added a series of notes, which included a example of what a program might be like. We have no evidence that she wrote this conceptual program herself, and even if she did it didn't make her the machine's programmer.
The claim that King wrote programs comes up again in Matt Parker's entertaining Things to Make and Do in the Fourth Dimension, which I'm currently reading for review. But of more interest is his description of building a working computer (admittedly only capable of adding up to 16) with 10,000 dominos by using the interaction of falling dominos to produce gates. This was a wonderful feat for which this tireless maths enthusiast should be congratulated. You can see the 10,000 domino computer in action below.
* I prefer Ada King to the more commonly used Ada Lovelace, though I admit I seem to be about the only one who does. Her full name was Ada King, Countess of Lovelace. While in principle a countess can be referred to by her title in place of surname, the usual reporting standard is to use the surname. So, for instance, when referring to the Duke of Bedford, he is called Andrew Russell, not Andrew Bedford. People sometimes get confused because the royals don't really have surnames, so there's no other choice with them. But I think with Ada it's primarily done because 'Lovelace' sounds more exotic.