Plan 28

2010 · Analytical Engine, In Progress · Work 519 of 1,871

Plan 28, 2010, Analytical Engine, In Progress

A modern workroom, a large flatbed scanner with a nineteenth century engineering drawing lying on its glass, a screen beside it showing the same mechanism as a rendered three dimensional model of gear columns in grey, the original ink drawing weighted flat on the table beside the screen, the table top plain, the scanner lid raised, a keyboard beside the screen, grey surfaces throughout, even flat light from directly overhead.

What it was used for
Plan 28 is not a device but an undertaking: to build Babbage's Analytical Engine, which never got beyond drawings. The programmer John Graham-Cumming proposed it on 21 September 2010 and has worked since 2010 with Doron Swade, who led the construction of Difference Engine No. 2; the Computer Conservation Society and the Science Museum in London are involved. The work has three stages: first digitise Babbage's archive, then study it academically, then build. So far the second stage is under way. It includes a searchable, cross-referenced catalogue of the drawings and a reading of the notebooks - on the account reported by i-programmer, seven to eight thousand manuscript sheets in some twenty-five volumes. The undertaking is named after one of Babbage's numbered general plans.
What was new
What is new is not the idea of rebuilding a Babbage machine - that was done in 1991. What is new is what this undertaking has to build first: not the machine but the understanding. For Difference Engine No. 2 there was a closed set of drawings. For the Analytical Engine there are hundreds of sheets from decades of work, showing successive and different designs. Swade's point is that a coherent and consistent understanding of the machine cannot be recovered from the mechanical drawings alone. That is the barb: the first reconstruction proved that a Babbage design holds; the second undertaking has shown that the design does not exist as one thing. Sixteen years after the proposal the machine does not exist; in early 2025 the project reported first exploratory trial constructions of individual mechanisms. The target named by i-programmer in 2018 was the 2030s, two hundred years after the design.

What the sources leave open: Thinly sourced. The state of play comes from the founder's 2011 blog, an i-programmer report of 27 January 2018 and Swade's project reports on plan28.org up to February 2025. Charity status and registration number were not checked. Swade's statement appears verbatim in his 2016 report, quoted by i-programmer; it is paraphrased here. According to i-programmer Plan 28 is Babbage's most complete design; Babbage himself gave that name to Plan 13 in 1849.

Further reading