by Thomas Smith, for Mrs. Gedney
Sun. 20 August 1881
My dinner with Ms. Billings was very pleasant. The main topic of discussion was the Census. I am involved in the Census of 1880, which is still going on. Dr. Billings, Kate's father, commented that there should be a way to automate the tabulation of Census statistics. I have seized upon this idea. I know how! My design will require some work on proving concepts, but I am sure that it will work in the end. I gave a short description of my idea at
dinner. Kate seemed quite impressed... this night has been a dream for me. It is unfortunately rare to meet a woman who appreciates electrical and mechanical engineering.
My design is this: a machine will sense holes in a paper tape to advance counters. The machine must be capable of counting combinations of holes; for instance, it must be able to increment different counters depending on whether someone is colored and female, white and female, colored and male, or white and male. Each of those categories should have only one counter. The real goal is to generalize this so that combinations of arbitrary complexity may be counted.
How will the counters work? It is my thought that they will sense an electrical impulse and increment. This would happen by the force of an electromagnet. It would be kept from moving too far by a ratchet wheel, or some similar device. A wheel design would make it hard to implement with electricity. Perhaps some analogue on a flat surface?
How will it sense holes? I am thinking about an electrical design so that, where there is a hole, an electrical contact is made. This could be implemented using some brushes and a metal base under the tape. The base would be charged and the brushes would be connected to the counters. This could be a problem because of the spacing of records--if the brush hairs wore down or detatched, they
could complete a circuit where none should be completed, causing an error. Another approach would be to use metal pins, and an intermittent action. The brushes would work best when the tape kept moving, but pins would get caught on the edges of holes.
(added months later, after reading through the above entry):
Intermittent action... that brings an idea to my mind. Perhaps a paper tape is not the best medium. I don't like it anymore. The problem was that if, for example, you wanted any statistics regarding Chinamen, you would have to run miles of paper to count a few Chinamen (Austrian 14). I am reminded of a technique used on the railroad: a punch-photograph. Each person has a card, with notches cut on the edges denoting certain traits, like big nose, brown hair, blue eyes, et cetera. Each card represents one person.
It's quite simple! The cards can be sorted, eliminating the
problem. Wonderful!
On a somewhat less happy note, Kate has married someone else. I am working on getting over this.
Notes:
This is a genre about Hollerith's thought process. I think that it is substance, actually, contrary to my previous notes on the piece, which said that it was not because it did not fulfill my original goal. This is one of my better pieces. When I have been able to write pieces of substance, I have gotten ambitious and overestimated my ability to write more substantive pieces. Therefore, most of my pieces of substance were originally labeled as not being substance, and vice versa. This diary entry represents
the thought process which might have been behind Hollerith's
realization about the use of cards.