The Greatest Machine Never Built 132
mikejuk writes "John Graham-Cumming is the leading light behind a project to actually build the analytical engine dreamed of by Charles Babbage. There is a tendency to think that everything that Babbage thought up was little more than a calculating machine, but as the video makes 100% clear the analytical engine was a real computer that could run programs. From the article: 'Of course Ada Lovelace was the first programmer, but more importantly her work with Babbage took the analytical engine from the realms of mathematical table construction into the wider world of non-mathematical programming. Her notes indicate that had the machine been built there is no question that it would have been exploited just as we use silicon-based machines today.
To see the machine built and running programs would be the final proof that Babbage really did invent the general purpose computer in the age of the steam engine.'"
Re:Whose opinions are they refuting? (Score:5, Informative)
Yeah, I'm not quite sure where they got that from, unless it's based on popular confusion with the Difference Engine, an earlier design that could not do general-purpose, programmable computation.
Babbage as a forerunner of modern computing isn't a recent acknowledgement either: many of the digital-computing pioneers explicitly referenced him, and compared their work to his, usually viewing his work favorably and chalking up its failures to practical implementation problems, not severe drawbacks in the design. Here's [google.com] a 1958 article in New Scientist crediting Babbage, which even includes a table comparing the Analytical Engine with EDSAC [wikipedia.org].
The only serious controversy I know of is whether the design could've been built with technology of the time, not whether the design itself was sound. See e.g. this 1998 journal article [univr.it], particularly p. 34 (6th page of the PDF), which concludes that it could probably have been built, though it would've been quite expensive and required the top machining abilities of the day.
Re:erm.. it was built (Score:5, Informative)
You're probably thinking of the Difference Engine that the London Science Museum built in 1991 [sciencemuseum.org.uk] (output mechanism added in 2000). Afaik nobody's constructed an Analytical Engine, which is considerably more complex to build.
Re:erm.. it was built (Score:4, Informative)
Re:What if... (Score:4, Informative)
For those who want to learn about that book instead of just buying it, here's a better link. [wikipedia.org]
Re:I'm not sure it is even a controversy any more (Score:4, Informative)
Look at the other research the Admiralty funded back then - timekeeping (pocket watches), astronomical calculatons (octants, sextants and easy to use calculation tables), tide calculstions, leading to signal processing and Fourier transforms, fluid dynamics and Navier Stokes equations.
Re:Great idea, probably not happening (Score:5, Informative)
You are correct that I care about the PR side of things. I need to because I need to raise a substantial amount of money.
But it's far from all PR. There's now a registered British charity with a board of trustees and the pre-eminent Babbage expert, Doron Swade, who built the Difference Engine No. 2 at the Science Museum is running the technical side of the project.
Study of the digitized plans has been underway since February and some first results will be announced this summer. We actively want to build a 3D working model in a tool like Autodesk.
Re:wrong (Score:4, Informative)
Yes, there is a need to describe a device as Turing complete if you are going to be accurate. The parent post above got it dead right when he said that Konrad Zuse didn't even get the right millennium in terms of when the first (known) mechanical computer was built. That would likely be the Antikythera mechanism [wikipedia.org], and there is reason to believe there were machines of similar complexity which existed earlier. The Egyptians and Babylonians had clockworks which predated even this device by several millennia, and depending on your definition you could even describe things like Stonehenge to be a computer.
The point of suggesting Turing completeness is that you get into the realm of programmable computers that have a set of characteristics all to themselves... and that Alan Turing described mathematically a set of characteristics that distinguishes these early mechanical computers (not merely calculating devices either) from more modern computational machines that are typically called "computers" in a modern context.
The issue with Konrad Zuse is that he did build the first functioning Turing complete device (well... made a strong attempt at it with the Z1), with other computers being built about the same time.
The amazing thing about the Babbage design is that it was using technology of a much earlier era and still could get the job done. That is why it is an amazing design and why so many "what if" statements keep getting made about it. Britain certainly had accurate tables for performing firing solutions with their weapons, but just imaging what World War I would have been like with a United Kingdom having at its disposal nearly a half century of computer technology experience is certainly something that would have changed the outcomes of war. Or perhaps if like many inventions the Germans took designs and ideas from the British and refined them to a much larger degree... imagine what Imperial Germany could have done with the same machine and subsequent designs incorporating the electrical circuits that existed in the early 20th Century.