10.3 Object-Oriented Programming

Object-oriented programming is a style of programming where programs are broken down into objects that can be combined to solve a problem or model a simulated world. The notion of designing programs around object manipulations goes back at least to Ada (see the quote at the end if Chapter 6), but started in earnest in the early 1960s.

During World War II, the US Navy began to consider the possibility of building a airplane simulator for training pilots and aiding aircraft designers. At the time, pilots trained in mechanical simulators that were custom designed for particular airplanes. The Navy wanted a simulator that could be used for multiple airplanes and could accurately model the aerodynamics of different airplanes.

Project Whirlwind was started at MIT to build the simulator. The initial plans called for an analog computer which would need to be manually reconfigured to change the aerodynamics model to a different airplane. Jay Forrester learned about emerging projects to build digital computers, including ENIAC which became operational in 1946, and realized that building a programmable digital computer would enable a much more flexible and powerful simulator, as well as a machine that could be used for many other purposes.

Before Whirlwind, all digital computers operated as batch processors where a programmer creates a program (typically described using a stack of punch cards) and submits it to the computer. A computer operator would set up the computer to run the program, after which it would run and (hopefully) produce a result. A flight simulator, though, requires direct interaction between a human user and the computer.

The first Whirlwind computer was designed in 1947 and operational by 1950. It was the first interactive programmable digital computer. Producing a machine that could perform the complex computations needed for a flight simulator fast enough to be used interactively required much faster and more reliable memory that was possible with available technologies based on storing electrostatic charges in vacuum tubes. Jay Forrester invented a much faster memory based known as magnetic-core memory. Magnetic-core memory stores a bit using magnetic polarity.

The interactiveness of the Whirlwind computer opened up many new possibilities for computing. Shortly after the first Whirlwind computer, Ken Olson led an effort to build a version of the computer using transistors. The successor to this machine became the TX-2, and Ken Olsen went on to found Digital Equipment Corporation (DEC) which pioneered the widespread use of moderately priced computers in science and industry. DEC was very successful in the 1970s and 1980s, but suffered a long decline before eventually being bought by Compaq.

Ivan Sutherland, then a graduate student at MIT, had an opportunity to use the TX-2 machine. He developed a program called Sketchpad that was the first program to have an interactive graphical interface. Sketchpad allowed users to draw and manipulate objects on the screen using a light pen. It was designed around objects and operations on those objects:1

In the process of making the Sketchpad system operate, a few very general functions were developed which make no reference at all to the specific types of entities on which they operate. These general functions give the Sketchpad system the ability to operate on a wide range of problems. The motivation for making the functions as general as possible came from the desire to get as much result as possible from the programming effort involved$\ldots$ Each of the general functions implemented in the Sketchpad system abstracts, in some sense, some common property of pictures independent of the specific subject matter of the pictures themselves.

Sketchpad was a great influence on Douglas Engelbart who developed a research program around a vision of using computers interactively to enhance human intellect. In what has become known as "the mother of all demos", Engelbart and his colleagues demonstrated a networked, graphical, interactive computing system to the general public for the first time in 1968. With Bill English, Engelbard also invented the computer mouse.

Sketchpad also influenced Alan Kay in developing object-oriented programming. The first language to include support for objects was the Simula programming language, developed in Norway in the 1960s by Kristen Nygaard and Ole Johan Dahl. Simula was designed as a language for implementing simulations. It provided mechanisms for packaging data and procedures, and for implementing subclasses using inheritance.

In 1966, Alan Kay entered graduate school at the University of Utah, where Ivan Sutherland was then a professor. Here's how he describes his first assignment:

Head whirling, I found my desk. On it was a pile of tapes and listings, and a note: "This is the Algol for the 1108. It doesn't work. Please make it work.'' The latest graduate student gets the latest dirty task. The documentation was incomprehensible. Supposedly, this was the Case-Western Reserve 1107 Algol---but it had been doctored to make a language called Simula; the documentation read like Norwegian transliterated into English, which in fact it was. There were uses of words like activity and process that didn't seem to coincide with normal English usage. Finally, another graduate student and I unrolled the program listing 80 feet down the hall and crawled over it yelling discoveries to each other. The weirdest part was the storage allocator, which did not obey a stack discipline as was usual for Algol. A few days later, that provided the clue. What Simula was allocating were structures very much like the instances of Sketchpad. There were descriptions that acted like masters and they could create instances, each of which was an independent entity. $\ldots$

This was the big hit, and I've not been the same since$\ldots$ For the first time I thought of the whole as the entire computer and wondered why anyone would want to divide it up into weaker things called data structures and procedures. Why not divide it up into little computers, as time sharing was starting to? But not in dozens. Why not thousands of them, each simulating a useful structure?

Alan Kay went on to design the language Smalltalk, which became the first widely used object-oriented language. Smalltalk was developed as part of a project at XEROX's Palo Alto Research Center to develop a hand-held computer that could be used as a learning environment by children.

In Smalltalk, everything is an object, and all computation is done by sending messages to objects. For example, in Smalltalk one computes (+ 1 2) by sending the message + 2 to the object 1. Here is Smalltalk code for implementing a counter object:

class name counter
   instance variable names count
   new count <- 0
   next count <- count + 1
   current ^ count

The new method is a constructor analogous to make-counter. The count instance variable stores the current value of the counter, and the next method updates the counter value by sending the message + 1 to the count object.

Nearly all widely-used languages today provide built-in support for some form of object-oriented programming. For example, here is how a counter object could be defined in Python:

class counter:
   def __init__(self): self._count = 0
   def rest(self): self._count = 0
   def next(self): self._count = self._count + 1
   def current(self): return self._count

The constructor is named __init__. Similarly to the object system we developed for Scheme, each method takes the self object as its parameter.

  1. Ivan Sutherland, Sketchpad: a Man-Machine Graphical Communication System, 1963