Last week, the ACM announced this year's Turing Award would go to Pat Hanrahan and Ed Catmull for their work on computer graphics. The Turing Award is often colloquially regarded as the "Nobel Prize in Computer Science." It also comes with a cool $1M (not far off the 8M Swedish Krone you get for a real Nobel Prize). If you play video games or watch movies with CGI, then you are looking at a lot of their work. If you have kids, and probably even if you do not, you will have seen some of Pixar's movies such as Toy Story. Ed Catmull was the CTO at Pixar, and Pat Hanrahan was one of his first hires.

I've written about previous honorees in past years although I have recently discovered that I got the years wrong. This year's award, despite it being March 2020, is officially the 2019 Turing Award. You can take one off the years in the links below, too:

• Geoff Hinton, Yann LeCun, and Yoshua Bengio Win 2019 Turing Award (for getting neural network training working)
• Hennessy and Patterson Receive the 2018 Turing Award (for RISC processors)
• The World Wide Web Was Born 30 Years Ago This Month (Tim Berners-Lee for the World Wide Web)
• Turing Award; Google's First Crash (Whitfield Diffie and Martin Helman, for public-key cryptography)
• The Most Important Operating System Ever and Why You Shouldn't Trust Ken Thompson (Ritchie and Thompson, for Unix)

Catmull, Hanrahan, and Pixar

Pixar had its roots in LucasFilms' division Industrial Light & Magic. Lucas hired Ed Catmull to create the Computer Graphics Division in 1979. I remember going to DAC in about 1984 and they were there showing the most amazing graphics that I'd ever seen, I think running on hardware that they had created. Silicon Graphics, which was probably the most influential computer graphics company in that era, was not founded until a couple of years later in 1981. Lucas hired Catmull to run the division.

In 1986, LucasFilm decided that building graphics hardware wasn't really core to their business, and they sold (most of) it to a guy you've probably heard of, Steve Jobs. At that date, Jobs had left Apple but not yet returned, and was CEO of NeXT Computer. The spun-out division was renamed Pixar Animation Studios. One of Catmull's first actions after Pixar came into existence was to hire Pat Hanrahan who he had worked with before.

I worked in computer graphics back when it was very primitive, in the early 1970s. Between high school and university, I worked for my godfather for several months. He was a professor (lecturer in Brit-speak) in the Cambridge University Engineering Department working on (mechanical) computer-aided design. In those days, computers were not powerful enough to do anything close to what they can do today. Mechanical design was done using wire-frame models and the big problem to be solved was called "hidden line removal"—how to remove the lines that actually could not be seen because the object obscured them. It was too computationally complex in all except trivial cases. In fact, the solution would turn out to be not just more powerful computers (thank you Moore's Law) but something called Z-buffering, sorting all the polygons in the image from back to front. Oversimplifying, if they were drawn in that order then the final image would only have the ones you could see from the front. You could further optimize it by removing any polygon that pointed away from the viewer, and any others that were obscured completely. The technique was pioneered by...yes...Ed Catmull.

Pixar's movies are all powered by rendering technology called RenderMan. Pixar put the specification language in the public domain but kept their own implementation proprietary. That's how Pixar movies looked so much better than anyone else's, especially in the early days. RenderMan was created by Pat Hanrahan and it allowed complex lighting, shade, and shadows to be rendered reasonably realistically using ray tracing. He also invented texture mapping.

The contributions of Catmull and Hanrahan are fundamental to what makes modern 3D computer graphics work. A lot of what GPUs do is accelerate these basic algorithms so that modern video games on consoles and smartphones are close to photorealistic.

The Genesis of Computer Graphics

Computer graphics were pretty much invented by Ivan Sutherland in the 1960s with Sketchpad, the topic of his PhD dissertation. It was the first program to use a graphical user interface (as opposed to a keyboard). He would go on to work with David Evans at the University of Utah. They would also form Evans & Sutherland, a computer graphics company focused on building flight (and other vehicle) simulators for commercial and military use. In the early 1980s, they realized that to do this effectively going forward, they would need to design their own chips and so they were one of the initial investors when an ASIC company got started in Silicon Valley called VLSI Technology. A year later VLSI would hire me.

Computer Graphics Today

Ever since then, computer graphics has required specialized silicon. First, Jim Clark at SGI built a whole company on accelerating graphics algorithms in silicon. Now, nobody builds special graphics workstations, they just use GPUs to go alongside a server or a smartphone. The silicon design moved to companies like NVIDIA, AMD (ATI originally), Intel, Imagination, and Arm.

More significantly for this post, in the 1970s, Ivan Sutherland had a PhD student...yes...Ed Catmull. So there is a direct line from Sutherland, often called "the father of computer graphics", through Ed Catmull, to Pixar: Toy Story, Ratatouille, Finding Nemo, and not just more but more to come. But it goes beyond that, since these approaches to computer graphics, and RenderMan in particular, are the heart of most CGI used in many non-animated movies and TV programs.

The world, especially on screens, would look very different today without Pat Hanrahan and Ed Catmull.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.