Information Technology Reference
In-Depth Information
into the machine and then return the output to the users when the jobs were
completed. The “operating system” was just the loading routine used by the
operator to schedule the jobs on the computer plus the collection of I/O subrou-
tines. As commercial computers began to appear in the early 1950s, such batch
processing was the norm.
By the mid- to late 1950s, the limitations of batch processing were becom-
ing apparent, and in universities there was a great deal of experimentation
with the idea of more interactive computing. In 1955, John McCarthy ( B.3.7 ),
one of the pioneers of artificial intelligence , spent a summer at IBM's laboratory
in Poughkeepsie and got to learn computer programming through batch pro-
cessing on the IBM 704 computer. He was appalled at having to wait to learn
whether or not his program had run correctly. He wanted the ability to debug
the program interactively in “real time,” before he had lost his train of thought.
Because computers were very expensive systems at that time, McCarthy con-
ceived of many users sharing the same computer at one time instead of just
being allowed access to the machine sequentially, as in batch processing. For
such sharing to be possible, multiple users had to be connected to the machine
simultaneously and be assigned their own protected part of memory for their
programs and data. Although there was only one CPU, each user would have the
illusion that he or she had sole access to it. McCarthy's idea was that because
the computer cycles from instruction to instruction very quickly on a human
timescale, why not let the CPU switch from one memory area and program to
another memory area and program every few cycles? This way the user would
have the illusion that they are sole user of the machine. He called his concept
time sharing :
B.3.7. John McCarthy (1927-2011)
contributed many groundbreaking
ideas to computing, such as time
sharing, the LISP programming
language, and artificial intelligence.
In recognition of his pioneering
work in computer science, he
received the Turing Award in 1971.
Time-sharing to me was one of these ideas that seemed quite inevitable.
When I was first learning about computers, I [thought] that even if [time
sharing] wasn't the way it was already done, surely it must be what everybody
had in mind to do. 13
Time sharing was not what IBM had in mind. It is perhaps understandable that
IBM had little interest in time sharing and interactive computing, despite its
longtime involvement in postwar projects with MIT, because all of its business
customers were happy with their new batch-mode IBM computers. In order to
implement time sharing, McCarthy needed IBM to make a modification to the
hardware of the 704. This was needed for an “interrupt” system that would
allow the machine to suspend one job and switch to another. Fortunately IBM
had created such an interrupt modification for the Boeing Company to connect
its 704 computer directly to data from wind-tunnel experiments. IBM allowed
MIT to have free use of the package and in 1959 McCarthy was able to demon-
strate an IBM 704 computer executing some of his own code between batch
jobs. In his live demonstration, the time-sharing software was working fine
until his program unexpectedly ran out of memory. The machine then printed
out the error message:
B.3.8. Fernando Corbató pioneered
the development of operating
systems that allowed multitasking
and time sharing. He stated a rule of
computer science called Corbató's
law, according to which “The number
of lines of code a programmer can
write in a fixed period of time is the
same independent of the language
used.”
THE GARBAGE COLLECTOR HAS BEEN CALLED. SOME INTERESTING
STATISTICS ARE AS FOLLOWS . . . 14
Search WWH ::




Custom Search