The cards contained a snippet of code probably written in Fortran or Cobol—ancestors of sleeker languages like today’s Java or Swift. I manually entered 76,764 characters into a plaintext file, hoping to somehow convert the punched markings into something meaningful. This seemed unlikely because the text stamped on each card was an incomprehensible heap of letters and numbers.

I double-checked my translations, and became intimately familiar with the cards as I carefully examined and stacked them: the bent corner of card 657, the smudged ink on 618 that made me imagine a clerk unjamming a machine at the end of the day. These observations led to the discovery of more hand-written notes throughout the deck. Some numbered cards confirming my schema and some thoughts apparently jotted down on whatever was at hand.

I scanned these snippets of handwriting, and began searching online for a way to identify their author. The obvious place to begin was ARPA, now DARPA, back then the Advanced Research Project Agency, whose logo was stamped on each card. I downloaded a program from Github that could index an image library and search for handwriting matches, so I hunted for any public records from ARPA that included handwriting. The National Archive has a searchable collection, where I found scattered photographs and documents containing signatures and other digitized scribbles. The motherload was a three hundred page PDF of employment records.

There was an awful lot of Photoshopping and file wrangling, but eventually I got the software’s parameters just right and the pulsing progress bar disappeared, rewarding me... with a name.

Moira Duncan started working at ARPA in 1959 as an analyst for research and development. She had graduated from the City University of New York the previous year, her apparent brilliance in information theory attracting the attention of early computing agencies. She moved away from her family in Brooklyn without much fanfare, presumably excited about having access to the cutting-edge computers ARPA was developing in Arlington.

By all accounts, Duncan was a very inventive programmer and contributed to a number of projects at the Agency. I found some mention of her later years—when she was the resident computer expert of Arlington—frustrating her co-workers as she became moody and withdrawn, but no less brilliant. A list of ARPA projects from 1978 included several with Duncan listed as the sole assignee, including one mysteriously titled Elysium. She passed away in 1981, having watched the computer age shudder into existence but missing its arrival on the world stage.