The Wizard that Didn't Mode Me In

Larry Tesler got annoyed. Good. Somebody had to. Software made life harder than necessary on purpose–well, not on purpose, but it sure felt that way.

In a quiet corner of America, programmers were working tirelessly in front of an intimidating screen. The computer demanded its constant attention, cracking a proverbial whip. “Work harder, work harder”, and the programmers, as if on cue, “I will, I will!” Computers are users; humans are accessories.

A Computer only a Programmer could love

Imagine, if you will, at the helm of the famous editor Bravo. You are typing along. Then you typed ‘E’. The entire document was selected. You typed ‘D’. Gone. Then ‘I’, you are now in insert mode. Then you typed ‘T’. Now all you see is the letter T, rubbing your nose in dirt. Where did the document go? It’s not gone. Its gone gone.

In another, there was the oN Line System (NLS). You told the computer what to do first then what to do it to second. You want to add a word? Switch to “Insert Mode”. Delete? Switch to “Delete Mode”. Okay, but what if you do not remember which mode is currently active? You’re on your own. There was no graphic user interface to remind you. So, press “d”: you might add to your document or you might delete it. Plus, if you do delete it, you have yourself to blame, cause there is no undo.

A Tyranny of Modes

The problem? Modes. These computers ran on limited memory and processing speed. This meant that computers could only perform a certain number of tasks concurrently and efficiently; otherwise, the work would get queued and worked on serially. The programmers were strapped–-they needed to be creative. To avoid task degradation, the programmers wrote the software to help the computer focus its attention on a much smaller set of well-defined tasks at a given time. Modes kept the computer going without the overwhelm. The computer did its job. The programmers were happy. You weren’t even an afterthought. Not because they were being cruel. Because their user was the machine.

Back then, if you wanted to operate a computer, you could not go in with your quick wits and clever quips alone. You needed to be a certain kind of expert. Native. You had to think like the machine. you had to be the machine. Every action required memorizing commands. It was all on you. Thankless if you do it right, punishing if you do it wrong.

Tesler’s colleagues thought NLS was intuitive to use.

Intuitive? How, if you need to be vigilant in what you are doing, exact in every keystroke, and at the mercy of the computer if you errored? Certain, only if you knew. Computers were meant to be operated by the masses, not by the few who were slaved to be expert in its operation.

First Glimmer of a Solution

Tesler imagined an interactive page makeup system that would simplify the process.

He didn’t start from zero. He had seen the pieces.

From 1968 to 1970, Pentti Kanerva showed his PDP-10 port of Brian Tolliver’s full screen text editor, TVEDIT. Like the NLS and Bravo, this was also a modal editor. The biggest changes are what Kanerva implemented. The clipboard. The Oops. The two-step move. In other systems, once a text is deleted, it is gone forever. Kanerva’s clipboard caught it. The Oops command returned it. The two-step separated the cutting from the pasting. Finally, a user interface that considered you for once. Tesler saw this. He saw the pieces.

Now he needed to convince others.

In 1973, Tesler worked directly with Jeff Rulifson, the designer of NLS. In NLS, you told the computer what to do first then to what second. Tesler argued you should be able to select first, then act. That way, no harm done if you pick wrong. If you pick the wrong command, you will know what went wrong so you can fix it. In other words, the computer shouldn’t punish you for exploring.

Rulifson listened. He understood the logic, but it would require him to rebuild the interface from the ground up. He had a system that worked. He had users that were accustomed to it. He was not ready to abandon it. It wasn’t broken–not in the way that demanded fixing. Not until someone asked: broken for whom? He did not outright reject Tesler’s idea either. Other colleagues were more willing to listen to Tesler’s proposals, though.

If Rulifson did not build it, Tesler would. Tesler did this not because he was perfectionistic, but he strongly believed you deserved better. He used Smalltalk. It let him prototype quickly. He stripped away the modes. He built a simple text editor. Click to position the cursor, type to insert, select text, then act. No insert mode. No command mode. Just editing.

The Blank Screen Test

He wanted proof. So he found them–the ones the industry forgot. People who had never touched a computer. They were not programmers. They were not power users. They were just people. Tesler sat them in front of the blank screen. No instructions. No menus. He watched. They clicked. They typed. They made mistakes. They corrected those mistakes. It became easy. They did not need a manual. They did not need a training course. They just needed a tool that worked the way they expected.

Tesler had his proof.

After completing this work, Ginn and Company immediately put Tesler to work. Ginn needed two applications. One for galley editing and the other for arranging pages. Ginn recruited the help of Tim Mott. They wanted this launch without a hitch. Together, Tesler and Mott set out to develop Gypsy.

They did not start from scratch.

Back at the lab, several of the computers Tesler, Mott, and fellow researchers were using had already been running Bravo. They took the source code. They replaced the modal interface with a modeless one. The result felt different. Click to position cursor. Double-click to select a word. Drag to select text. Cut-and-paste. Copy-and-paste. Bold, italic, underline. Suddenly, editors felt seen and heard. They could proofread without fighting the software now.

Gypsy worked. The page layout application, which Tesler dubbed Cypress, did not. It was too slow. Xerox never shipped it. It would have to come later. But the text editor was ready.

The Five-Day Sprint

On July 17, 1980, Tesler started working for Apple. When he arrived, he was told the user interface design for Lisa was nearly complete, that they were going to write up the documentation, and finalize it. All decisions had to be made by July 23rd. 5 days. How could Tesler contribute? The following week, the team needed to prepare the write up. Certainly, he did not want to cause any delay, compromise on the quality, let alone jeopardize the operation. So many people’s blood, sweat, and tears were on the line. With 5 days to do it, how could he be sure he made the right choices? Tesler needed to strategize fast.

He worked with Bill Atkinson. Since Tesler did not have enough time to run a proper user study, they did the testing themselves. Bill Atkinson built the prototypes. Everyone tested them out. Everyone evaluated the prototypes for problems.

However, this approach also needed additional perspective. They needed to know what parts would be the easiest to train. They sought the opinion of Sue Espinosa, the corporate trainer. The group discussed all the issues. They took it to Steve Jobs to make the decision. Jobs told Tesler what he liked. Tesler took the feedback to the group for another round of focused testing. Tesler wrote a memo recommending a one-button mouse. He mapped X to Cut, C to Copy, V to Paste, and Z to Undo. Atkinson wrote the interface specification. The decisions held. Lisa shipped. The world changed.

The Law of Conservation of Complexity

Tesler had five days. He changed the future. Then the future spread. The Lisa. The Mac. Word. Windows. Your phone. Cut, copy, paste is everywhere. You use it without thinking. That is the point. Tesler did not make computers simple. He made the complexity invisible. Like a wizard who respects his audience. They call it the Law of Conservation Complexity. I call it common decency. More importantly, as impressive as his feats were, Tesler found the magic in the tools that were given to him. He was a wizard, a darned good one. No less than you or me.

The Wizard that Didn’t Mode Me In