Someday, when the robots eventually take over and decide to write their own history books, they will devote an entire chapter to Willow Garage.
The organization was central to robotics progress over the past twenty years, even though its role is relatively unknown outside of technical circles. And even there, the story of how they succeeded in creating a common Robot Operating System (ROS) is not well understood.1 The lessons — especially their strategy to embolden the following generation — are applicable to other aspiring scenes.
Willow Garage was technically a company, but it operated as (and amounted to) a well-heeled amateur scene. It was started by Scott Hassan in 2006. Hassan was a computer scientist and entrepreneur who had already made it big. He had started a company and sold it to Yahoo. He was also a functional co-founder of Google, having helped Larry Page and Sergey Brin write the original code for the search engine as well as invested in the nascent company. Financially speaking, he was set.
His fortune was an important factor in being able to start Willow Garage, but so was his motivation and philosophy. Hassan recognized the important role that open-source software had played in his business success. By building on the stack of open tools, he was able to quickly get his software company, eGroups, off the ground and serving customers without having to build everything from scratch. He theorized that robotics needed a similar boost — contributions at the core operating system layer that would make building applications easier for everyone.
He started with a building and a budget. The building was a nondescript single-story office complex in suburban Palo Alto, not far from the Stanford campus and the SRI research facility. The budget was enough to support up to 60 people and researchers, which he recruited carefully and doggedly.
Two of his first recruits were Eric Berger and Keenan Wyrobek, graduate students at Stanford who were finishing their PhDs in robotics and leading the Personal Robotics Program there. When Hassan approached them, they expressed skepticism about the plan, mostly because “it didn’t fit any natural mental model” as Berger would say. The novel organizational structure and circumstances of the funding arrangement were unlike anything they had come across in academia. Ultimately, they rightly recognized and seized the opportunity. The money was good, but the mission was better. As students, the pair had experienced firsthand the barriers holding back the field. The nature of robotics required diverse skillsets: hardware and mechanical expertise, electronics, as well as a comprehensive set of software skills like path planning and computer vision. They saw the same opportunity that Hassan had identified: an open-source scaffolding for robotics would be critical for progress.
Hassan continued to pluck the best and brightest from Stanford, SRI, and beyond. Initially, the team was set to work on two projects: an autonomous car and an autonomous boat. When those efforts stalled, Hassan put the focus back on Berger and Wyrobek’s Personal Robotics project. While at Stanford, they developed PR1 as a hardware platform for the open software component. Willow Garage, with all its resources and talent, was now going to build the PR2.2 They set technical milestones, like navigating the building and plugging itself into a wall outlet, as well as manufacturing goals, which wound up being twenty machines; ten to be donated to research universities and ten to keep around the office. ROS, the Robot Operating System, was decided on as the software to run the PR2.
At the time, ROS was still mostly a Stanford project, having been started by Berger and Wyrobek and still leaning heavily on the open-source robotics work done by Morgan Quigley, another Ph.D. student at Stanford who was working in the Stanford Artificial Intelligence Laboratory. Eventually, an effort was made to formally merge various efforts to focus on ROS. Quigley would later comment that the fact that the project was between organizations — separated by both institutional boundaries and physical distance — helped because “it forced you to have a little more formality baked into the software to allow you to work on different components in parallel.”
As PR2 developed and Willow Garage grew, ROS development centered there. Quigley and the team published the ROS white paper and pushed version 1.0 on SourceForge. In addition to ROS, the PR2 became the other focus for the team, and the PR2 beta program — developing and delivering the common hardware to academic labs around the country — was the main strategy for gaining adoption of ROS amongst developers. However, it was another initiative at Willow Garage that would prove more effective: the internship program.
The internship program at Willow Garage, led by Melonee Wise, was legendary. The story, in her own words:
When we started really working on the Personal Robots Program and on ROS, I went to Steve and said, “The way that you make companies grow, and the way that you get people excited about what you’re doing, is to have a really big internship program. If you want to make sure that people start using ROS, the best way to do it is to bring people here and make them use ROS, and then send them all back to their labs.” I knew this because I’d done a lot of internships, and it was obvious to me how they were successful. Internships were about selling the Kool-Aid.
The internship program made a huge impact on robotics. It wasn’t just about learning stuff; I made a real effort to make their entire time at Willow awesome. We had parties, we went rafting, wine tasting, laser tag … And that’s important, if you want to create this connection. We had jackets. I know it’s stupid, but we had to make sure that when they went back, everyone saw it. People called it the Willow Mafia.
It worked. Not only was Willow Garage attracting all sorts of new talent to the team, but they were sending back those talented students to their respective academic labs, and with ROS freshly installed into their working memory. When they got back to their Universities, they were biased towards using ROS on whatever robotics project their team was trying to concoct. Unexpectedly, ROS was spreading out into the world before the common hardware platform had been shipped. Ken Conley credits the internship program:
Originally, we thought that the vehicle for ROS was going to be the PR2, but what happened, and this is really a credit to Melonee and what she did with the Willow Garage internship program, was that because we had so many interns come through from labs that had robots, it was the case that ROS was already running on lots of different lab robots at many different universities before we ever officially released the PR2.
A youth movement can buoy an amateur scene. Almost no one thinks about design on a generational scale, but it’s a powerful way to shape the future. Get the kids comfortable with the new tools and ideas, then wait for them to take over the world.
Companies and organizations often think about internships as a way to get cheap labor, or as a way to find and evaluate talent. This kind of program usually gets bucketed into the “education” or “recruiting” category and is often outsourced to schools or human resource departments. Wise and Willow Garage took an expansive view of the idea and got meaningful results.
Youth movements on their own aren’t usually sufficient to kickstart a scene — they must be connected to other ingredients like catalogers, kits, or open standards.
We saw a version in the underwater robotics world. As a high-school student, my teammate Eric Stackpole participated in the international MATE competition, a program where teams of students would design, build, and deploy a remotely operated vehicle (ROV) to achieve some required tasks. The experience was transformative for him. It exposed him to the world of underwater robotics and helped build his confidence in his emerging design skills. The event series has been popular among students around the world, as well as appreciated by marine technology companies who are eager to capitalize on talent development. At OpenROV, we hired several MATE graduates and always found them resourceful and knowledgeable. It’s a great program, but not an amateur technology scene (as I’ve defined them).
Every student group that came to the event started at square one. While this was probably good from an educational standpoint — everyone has to learn everything from scratch — it held the group back from changing the technological status quo.3
Contrast that with iGEM, the international competition where students design and build genetically engineered microbes. In the early 2000s, when Tom Knight, Randy Rettberg, and Drew Endy were puzzling about how to turn biology from a basic science to an applied engineering field, they turned their attention to the meta techniques of field-building. Biology was hard to engineer, but it could be made easier, they figured, if more minds started working on the problem and everyone was sharing their work — learning by everyone doing. The iGEM jamboree was a “coopetition” model where, every year, student teams compete to create the best new biological part or microbe. Each qualifying project is then added to the Registry of Standard Biological Parts, a growing database for the community to use and build on. iGem is a youth movement and an expanding biology kit — the combination has made it a major part of the growing synthetic biology scene.
Scientific publishing saw a similar generational change in the late 1800s with the emergence and transition to scientific journals. The uptake of the periodical format was not sudden or orderly. Nature started on the scientific periphery. The X Club was an elite club of scientists, known as “Darwin’s Bulldogs” for their aggressive espousal of Darwinian evolution. They helped lend credibility to the fledgling publication started by Norman Lockyer, but they viewed it largely as a mechanism to influence public opinion. It wasn’t considered high scholarly work, which was still the realm of the book and the deliberations of the learned societies. It wasn’t until a younger generation of scientists, those growing up in the shadow of the X Club, began to adopt the method of communication as a preferential publishing mechanism that things began to change. Nature became the venue and the winning benefit was speed. In this case, a youth movement combined with an accessible new communication medium permanently changed scientific publishing.
It can take a full generation (or several) for a new medium or tool to find its footing. Early adoption is a hint, but the true impact and format-defining work will come from the following generation — from the group that knows no different. In doing so, they re-route the path to the top.
Willow Garage took the long view and succeeded. They did it, in part, by building an effective youth movement. And their tools have spread well beyond academia. In 2019, a report showed that ROS was set to be installed on more than half of commercial robots being sold around the world. Achieving this level of ubiquity took foresight and patience on the part of the initial team. Ken Conley remembers the criticism at the time:
When we first started doing ROS, one of the things that was consistently said was that, “ROS is only for research robots, no one is going to use this in commercial robots.” We always knew at the time that it was more [than that], that we’d have to be patient for that. We knew that no one who had a current robotics company was going to throw away all of their software and switch to ROS. That would be an unrealistic expectation. But, what we were really trying to do was set a foundation for future growth in the robotics industry. I think we did a pretty good job of predicting that, and being ready for it.
If you want to change the world, equip the kids.
All of the quotes for this essay are from the oral history of ROS that was put together by IEEE: https://spectrum.ieee.org/wizards-of-ros-willow-garage-and-the-making-of-the-robot-operating-system
For those inclined, Clearpath Robotics recently made the design files for the PR2 available for download: https://clearpathrobotics.com/pr2-resources-2/
That’s changing thanks to Jill Zande’s leadership and foresight. She saw the impact that kits can have in advancing technology, and companies like Blue Robotics are making ever-better components for the teams to utilize. MATE will be an important part of the next great amateur scene in marine robotics, surely.