Frank Miller: Inventor of the One-Time Pad

Abstract

The invention of the one-time pad is generally credited to Gilbert S. Vernam and Joseph O. Mauborgne. We show that it was invented about 35 years earlier by a Sacramento banker named Frank Miller. We identify which Frank Miller it was, and speculate on what might have led him to his idea. We also discuss whether or not Mauborgne might have known of Miller's work, especially via his colleague Parker Hitt.

Keywords:

• Gilbert S. Vernum
• Joseph O. Mauborgne
• one-time pad
• Parker hit
• telegraph code

Acknowledgments

As always when we are doing historical research, we must first thank the generations of librarians and archivists who have collected, indexed, and preserved crucial documents over the decades. At least as important, they help me find obscure materials. We are especially indebted to the staffs of the Library of Congress (where we first found Miller's book), the Columbia University Library Network, the Stanford University Library, the California State Library, the Huntington Library, the New York Public Library, and René Stein of the library at the National Cryptologic Museum. Google Books deserves a special mention for its role in making many rare works easily available; they've scanned sources that we never would have found by conventional research, because there were no rational pointers to the works save for full-text searches. At our request, they also scanned a copy of Miller's book.

Betsy Rohaly Smoot of the NSA Center for Cryptologic History, an expert on Hitt, provided an extraordinary amount of help, checking and rechecking his diary and papers.

William Shurtleff's genealogy book was an extremely valuable resource; his private comments were also helpful.

We had limited time to travel; we were aided in “remote research” by Roy Frostig of Stanford University, who checked various newspaper archives there.

Craig Bauer and Drew Wicke of York College of Pennsylvania and John W. Dawson, Jr., retired from Penn State University at York, found Hitt's paper at the Army Signal School Technical Conference, confirming Kahn's opinion on when the observation on key length was made, but also showing the context of that observation.

Marianne Babal of Wells Fargo Historical Services provided important information on the Wells Fargo codebooks and superencipherment.

We have received a lot of useful information and guidance from many historians, including David W. Gaddy, and Dr. Edgar F. Raines, Jr. and Rebecca Robbins Raines of the U.S. Army Center of Military History. Norman Polmar of the U.S. Naval Institute and David Kahn also provided very important advice and pointers. Errors, of course, are all ours.

Notes

¹At my request, Google Books very kindly scanned a copy of Miller's codebook. The URL is listed in the bibliography entry.

²The first Wells, Fargo codebook dates to 1874; the Library of Congress has not been able to locate their copy, nor have we found another source. The Wells Fargo Historical Services office has codebooks dating only to 1877. Information from [22] suggests that in 1883 the test words were as much indicators to denote the length of the message. Still, if they were changed frequently enough they would serve as authenticators.

³[60] calls Miller a “founding trustee”. Arguably, this is correct, given the convoluted governance history of the university [18]; however, [6, p. 9] clearly identifies his predecessor in a given Board seat.

⁴Data retrieved from the University of Virginia Library Historical Census Browser, http://mapserver.lib.virginia.edu/

⁵The census record for the banker Frank Mills at https://www.familysearch.org/s/recordDetails/show?uri=http://pilot.familysearch.org/records/trk:/fsrs/rr_206956290/p_334654861&hash=HloWXpZgU9zB10k5M56iYku8TUc%253D shows a birth year of 1850, which is inconsistent with other sources and difficult to reconcile with Civil War service. The other data listed agree so well with [60] that we suspect an error in the record or in its digitization.

⁶Olcott himself had a very colorful career. After the war, he became a lawyer. He grew interested in spiritualism, converted to Buddhism, and was one of the founders of the American Theosophical Society. Most of the published information on him concentrates on his religious career.

⁷Kahn had speculated that Hitt formulated this precept before 1914. Research by Craig Bauer, Drew Wicke, and John W. Dawson, Jr., has confirmed this. However, the paper they found is by Hitt and Lieutenant Karl Truesdell, and does not acknowledge Mauborgne. Truesdell compiled the first set of multilingual frequency for the U.S. Army [38].

⁸One should not draw the conclusion that Hitt was a habitual plagiarist who stole from both Miller and Bazeries. Kahn quotes Hitt's memo as explicitly crediting Bazeries for the wheel cipher. Hitt was in fact punctilious about assigning proper credit [66].

⁹Neither Kahn nor his sources [37] provide any motivation for the move to one-time pads.