Only 40 years after Charles Babbage and
Ada Lovelace last shoveled coal into their
steam-powered 'difference engine' practical
Yankee Herman Hollerith built a machine
to tabulate the flood of data from the 1890
US Census. His data formats are still with
us. While Babbage and Lovelace had used
cards to program their engine to massage
'data' recorded on iron wheels, Hollerith
used paper cards to store the data.
Hollerith devised a 12 column system broken
into three 'zones' and nine 'number columns'
No punches was a zero, no zone punches meant
a number 1 to 9 and 27 alphas were coded
into three 9 letter 'zones'. If Holerith
had used binary, we might have had a 12
bit (16,0000) character set from the start-
enough for Chinese. But his primitive 'capitals
only' system became the data-storage standard
from the earliest days and was legitimized
by EIA standard RS-292. This standard is
augmented by ANSI X3.21-1967 governing the
holes in the card and ANSI X3.26-1980 governing
the use of the Hollerith code to encode
alphanumeric data on cards.
Hollerith's standards tended to keep data
in capitals and limited records 80 characters,
forcing many dates to be stored as two digits
and precipitating the Y2K crisis a century
later.
Long after the paper was gone, the record
and data format lived on in COBOL and mainframe
databases in the company Holerith founded
(IBM). Even today databases as ubiquitous
as the D&B credit histories are coded
within the limits established by Herman
Holerith in the 1880's. Ever buy a mailing
list and get it in ALL NASTY SCREAMING CAPITALS??
Herman did it. Ever get a download and find
that somebody works for the company Ge or
Gm? Blame Holerith plus an over simplistic
Holerith to ASCCI conversion (the Excel
'Proper()' Formula). Ever wonder why somebody's
record is truncated to 80 characters. Herman
is still at work..
