1 / 11
sketchpad
Fifty years ago today, Ivan Sutherland introduced the first graphical computer application: a drafting program called Sketchpad.
At a time when computers were operated with punch cards and command lines -- and the mouse had not yet been invented -- Sutherland used a light pen to manipulate lines and shapes on a screen. With Sketchpad, you could draw perfectly straight lines, change the size of shapes without altering the proportions, and create "rubber band lines" you could bend and stretch -- many of the things you can do today with programs like AutoCAD and Adobe Illustrator.
Sketchpad's legacy can't be overstated. It directly fed the creation of Douglas Englebart's oN Line System (the subject of the "Mother of All Demos"), which would influence, well, every graphical computing system that came after it, from the Xerox PARC Alto to the Apple Macintosh to the iPhone (see images above).
Sutherland says he was just doing what came naturally. "I was a graduate student and I was at the right place in the right time," he tells us. "I said I wanted to draw pictures, and management said that'd be fine." But Sketchpad was ahead of its time in more ways than one.
In building the application, Sutherland also created a predecessor to object-oriented programming, the model used today in programming languages such as Java, C++, and Objective-C. "It was the natural way to do graphics," he says. Later, at the University of Utah, he taught Alan Kay, creator of the SmallTalk language, which was a foundational influence on so many future programming languages.
Sutherland's father was a civil engineer and young Ivan learned to read blueprints by the time he was in high school. He later took some engineering courses as an undergraduate at Carnegie Tech (now Carnegie Mellon), but he was always frustrated by the process of drawing by hand, particularly how messy designs could get when you started erasing lines. The result was Sketchpad, which he started building in 1961 while working toward a Ph.D. at MIT.
"You'll notice that when you erase a line in Sketchpad it disappears entirely," he says.
Sketchpad is certainly his most famous creation (see video below), but in the years since, Sutherland has enjoyed a particularly fruitful career, spanning not only the graphics field but hardcore microchip design. After a stint at DARPA in the '60s, he worked at Harvard, where he developed the Cohen–Sutherland computer graphics line-clipping algorithm with Danny Cohen and invented the first head-mounted system for virtual and augmented reality with Bob Sproull. Then he moved on to the University of Utah, where he worked with Kay and several important graphics innovators, including Adobe co-founder John Warnock and Pixar co-founder and Disney president Edwin Catmull.
Sutherland later landed at Caltech as the founding head of the computer science department, and it was there that he fell in love with circuit design. "Carver Meade at Caltech had figured out how to design integrated circuits," he says. "It seemed more interesting than anything I had ever done, so I started doing it."
After CalTech, he co-founded a consulting company called Sutherland, Sproull and Associates with his brother Bert (an important computer scientist in his own right) and his former virtual reality collaborator Bob Sproull. According to Sutherland, this is when he made his most important contribution to science: logical effort, which enables chip designers to calculate how big a transistor should be used for optimal performance.
"Logical effort also tells you a lot of what you don't need to do," he says. The concept is now taught in electronics courses. Sun Microsystems later acquired the company to form the basis of Sun Laboratories, which still exists today as Oracle Labs.
These days he's a visiting scientist at Portland State University, where he and his wife Marly Roncken founded the Asynchronous Research Center (ARC) in 2009. "She's the important one. She's the director of the ARC," he says. "I just work here."
Together, Sutherland and Roncken research and teach asynchronous or self-timed design, which they think will result in more efficient chip designs. "Most chips today are synchronous -- they take a fixed time for every small step, like soldiers marching, Sutherland explains. "But one could easily do easy steps faster and take a longer time for harder steps. In many calculations, including addition, hard cases are rare." That leads to big waste both in terms of space and energy. In asynchronous design, chips can sacrifice a bit of speed when calculating hard problems but still solve easier problems quickly. He wrote an article on the topic in a recent issue of Communications of the ACM.
Having worked at several of the most important institutions in the history of computing, Sutherland has a few thoughts on what makes a successful research organization. First, you need an interesting problem to solve. Second, you need leadership, which he says is probably the most rare. And then you need money. He says that ARC has the first two, but he's worried about funding. Which is a shame because he's happy where he is.
"They keep asking me if I want to add the word professor to my title," he says. "But I like that title. I have the best job in the world."
http://www.youtube.com/embed/USyoT_Ha_bA
