The software is in the holes - The Computing Universe: A Journey through a Revolution - page 51

IF EMPLOYEE-HOURS IS GREATER THAN MAXIMUM

The language not only made the programs more intelligible to managers but
also provided a form of self-documentation that describes what the program is
supposed to do.
In May 1959 the U.S. Department of Defense started an initiative to develop
a common business language. This led to the COBOL ( Fig. 3.9 ) programming
language - for COmmon Business-Oriented Language - which was strongly
influenced by Hopper's earlier FLOW-MATIC language. For this reason, Hopper
is sometimes known as “the mother of COBOL.” What made the language so
successful was a declaration by the U.S. government a year later that it would
not lease or purchase any new computer without a COBOL compiler. At the end
of 1966, Hopper retired from the Navy with the rank of commander. Less than
a year later she was recalled to active duty and tasked with the job of rewrit-
ing the Navy's payroll system in COBOL. She was promoted to rear admiral
in 1985.
For the next twenty years, from about 1960 to about 1980, FORTRAN and
COBOL accounted for approximately 90 percent of all applications programs.
Backus went on to develop a notation to capture the “grammar” of a program-
ming language - that is, the way in which the special words and concepts of
a language can be put together. A Danish computer scientist, Peter Naur, then
simplified Backus's notation so that the grammar of any language could be cap-
tured in what is now known as Backus-Naur Form or BNF ( B.3.6 ). In the 1970s
Bell Labs produced a compiler-compiler , a program that could transform a BNF
specification into a compiler for that language. There has been much research
and experimentation with programming in the fifty years since FORTRAN and
COBOL. We will look at some of these developments in the next chapter.

Fig. 3.10. Early computers used punched
cards to input programs and data.

Early operating systems

In using even these early machines, it clearly made no sense for each
user to have to figure out independently how to interact with the computer.
Originally, users might input their programs and data - send instructions and
information to the computer - using a punched card or paper tape reader. Later
the input process might involve a keyboard, mouse, or, nowadays, a touch-
enabled tablet. Each user could also use disk drives to access and store data,
and could read off the results from a printer or some form of screen display. So
although the earliest computers had no real operating system - that is, no soft-
ware to control the operation of the entire computer system with anything like
the sophistication we see today - it was still useful to collect together all the
I/O subroutines - programs for input and output - and have them permanently
loaded on the machine.
In the earliest days of computers, users had to topic a time slot on the
machine, so graduate students were naturally allocated the nighttime slots!
Users loaded their programs into the machine using punched cards or paper
tape and then waited while the computer ran their program ( Fig. 3.10 ). This
personalized system quickly evolved to a more efficient system in which the
users were isolated from the machine by “operators.” Users now had to give
their program deck to the operator, who would load a batch of such programs

B.3.6. Peter Naur, a Danish computer
scientist, helped develop a successful
programming language called Algol
60. In 2005, Naur received the Turing
Award for his contributions to
computer science.