Increment: Did you always plan on a career as a developer?

Sarah Allen: Not at all. When I was 12, my mom brought home an Apple II. I spent a whole day learning to make [it] do things. It was delightful, [but] I bought into the stereotype that software engineers lived in small, windowless rooms, solving arcane problems. I had this idea that everybody wanted the world to be good, and it was just poorly implemented, and we all had to work at that. I didn’t see how computers would fit into that [work]. [When] I went to Brown, [I] was more interested in liberal-artsy things. I did painting and sculpture, and got a computer science degree as a backup. Though, over my career, I found that coding is much more creative than I ever expected it to be.

Did you realize that after you started CoSA?

CoSA created hypermedia on CD-ROM. We were competing with the likes of Time Warner, who could spend $2 million on a CD, not make any money, and [see the effort as] a good experiment. We took [our CD-ROM software], turning it into a .pix animation compiler. You could take audio, synchronize it with graphics, and play something off CD-ROM that was bigger than [what] would fit in [a computer’s] memory. We had 512 KB max. (I joke that engineers these days can’t count that small.) We also used to take turns being tech support. One day I was talking to somebody, and the software solved the problem he was facing. He said, “Wow, I didn’t think computers could do that.” I knew that when I was 12. I realized there was stuff [that was] obvious to me that other people didn’t think was possible—and that software comes out differently depending on who makes it. It’s a very creative act.

What does that look like in practice?

I did the design of the After Effects UI. The first designs looked a lot like Premiere, which came out around the same time. I don’t know why it seemed obvious to all of us that [representing] video in a computer would [mean] little thumbnails going across the screen horizontally. But a friend of a friend worked [at] a TV station, so we went and visited, and [we noticed] they didn’t have timelines. These screens would show you the whole composition. As an artist, I was like, “Of course.” One of the key innovations of After Effects was that we thought of time as going into the monitor, like the z-axis, when in math time is [usually] across the x-axis. Now you would call that human-centered design.

In 1992, you moved to Apple to work on the Open Collaboration Environment, an attempt to bring together a number of different Mac OS 7 communication apps under one networked umbrella, which was ultimately a dead end. What was that like?

Gursharan Sidhu had this vision [of a unified productivity suite] where on your desktop you’d have a directory of people [in a client app], Apple computers [would] interoperate across the network [using the AppleTalk protocol], and you’d have email and other applications [to communicate with the wider internet]. The vision was great [but], in 1993, HTTP was standardized. [Apple] decided that internet connectivity was a third-party opportunity, so [instead of open protocols] they had APIs. Two of my colleagues used to argue about what was going to win, HTTP or Gopher. They got blindsided, like many people, [by] what would happen with the web. I mean, I’m glad that “open” really means open source, open standards, open protocols.

Where did you go next?

After nine months [at Apple], I went back to CoSA, and three weeks after [that], [CoSA] was acquired by Aldus. Then Aldus was acquired by Adobe [in 1994]. I really didn’t want to be at such a giant company. Adobe was a couple thousand people, which is not really giant, but it felt that way to me. I worked on stuff nobody’s ever heard of, [like] video titling. We used the PostScript Print Engine to make really sumptuous overlays for video, and there was a really cool print driver, and we bundled them together and called it ScreenReady. That project was cancelled shortly after I left.

I wanted to learn about the web, and [computer] networking. Harry Chesley, who’d been at Apple, who I had learned a lot from, got a job at Macromedia [and I followed]. Macromedia was 250 people, which I thought was huge still! But I ended up on the Shockwave project, which was originally a code name. It was an interactive TV project—everybody thought that cable companies were all going to have set-top boxes that [would] allow you to interoperate with video through your remote. The whole idea was to build this stuff, except that the hardware that it was going to run on didn’t exist yet. The networks it was going to run on didn’t exist yet [either]. Everybody I’d ever known who’d worked on an iTV project had been laid off [within] a year or two, so I didn’t really want to work on [Shockwave]. But Harry had convinced [his boss] Norm [Meyrowitz] that we could prototype Shockwave on the internet. So one of my claims to fame is that I [only] agreed to work on [Shockwave] if I could work on the internet side. I joined Macromedia a month before they launched their [corporate] website, and in less than six months we [had] released the Shockwave Player.

That—along with further projects you led development on such as the Shockwave Multiuser Server, Flash Server, and Flash Player—made the web a much more colorful and interactive place, enabling new video, gaming, and chat sites. Did you foresee the explosion of multimedia that followed?

The idea that I could actually help people directly communicate with each other was really exciting to me, [but] in some ways I’d hoped for more. I’d had this idea that apps would allow people to do different things—grandparents [could] read their kids stories and make words come to life. But what the world actually “needed” was just Desperate Housewives and Lost. There were a lot of really creative things [made using Shockwave], but what saddens me today is that they’re not accessible anymore. I think it’s devastating that Adobe is just going to turn off the Flash Player. It’s like all the HyperCard stacks [from] the late ’80s, early ’90s—they’re all disappeared.

When did you realize that these were fading technologies?

I think the turning point was when Macromedia, and eventually Adobe, didn’t involve themselves in any of the emerging web standards. In the ’90s, many of the web-standards people were absolutely against having video on the web, other than as a file you [could] download. When we released Shockwave [in 1995], people said we were going to break the web.

There was this opportunity [when] HTML5 video started happening [as a result of new standards negotiations starting in 2004]. I think that, with the acquisition of Macromedia [in 2005], Adobe had acquired almost every company—aside from a team at Microsoft and at Apple—that really understood interactive multimedia. Almost all the engineers who had practical experience with this [area] worked at a company that did not involve itself in a substantive way with the [governing] web standards [first released in 2008].

When I was working on Shockwave [the original Macromedia Director plug-in for Netscape and Internet Explorer], I recall one executive saying to me: “I don’t understand why we need to be involved [with standards committees]. We didn’t do that for CD-ROM.” I didn’t have a response at the time, but later I thought about how we had to do all sorts of work-arounds for the fact that seek [times] were very slow on CD-ROM. Maybe if Macromedia had been involved in that standard, they would have ensured that it worked better for interactive content.

I think it set the industry back 10 years, and [led] ultimately to the demise of Flash. Proprietary technology was not the path. There’s a different way of doing business that drove the growth of the web. It’s counterintuitive, but by making software interoperable, by allowing people to export their data, by making it easy for people to leave, they are more likely to stay. It’s really about creating a commons. We share the expense of maintaining core infrastructure and libraries for interoperable file formats and protocols, which then increases the value delivered to customers.

How do you feel about the demise of these platforms, and the content that was published using them?

We need to preserve our history. I also think we have the opportunity to do something different. One of the things I learned about when we were doing Flash video was Doug Engelbart’s Augment [user interface] system. Kevin Lynch, who was [then] the CTO of Macromedia, had this brilliant idea of inviting Engelbart to virtually attend this launch and reenact his “Mother of All Demos.” Everybody else wanted to go to New York for the launch, but I was like, “Please, let me go and be Engelbart’s assistant and help him use our technology to participate.” In his house he still had the Augment system running in a simulator. He would read his email [on] it, he would save documents [on] it.

Living software is a completely different thing than preserved software. I’ve come to believe that the only software that survives and can possibly live a long life is open source. If I can’t export something into an open format then I’m very reluctant to use software. With the demise of Shockwave and Flash, After Effects is still going strong, but I feel that part of that is because the team that worked on it in 1990 still is working on it today. I think proprietary software inevitably dies. And so I wonder, what could I do? What piece of software, no matter how small, could I create that might outlive me?

Is that why you’ve spent much of the last decade involved in grassroots education?

Yeah. In the early part of my career, my wanting to make the world better was very separate from my making software. RailsBridge was a way I could use my expertise to help people, and, selfishly, not be in a room that was only guys.

What about your work as a Presidential Innovation Fellow? You helped spearhead the intragovernmental professional-development platform, Open Opportunities.

I was doing crowdsourcing at the Smithsonian, [but] the executive branch of the federal government had this problem: People are doing very similar things in different agencies, but struggle to collaborate because [by law] you need the secretary of your agency to say that it’s allowed [to work on projects not explicitly approved by Congress]. The idea [was to establish a microtask sharing site so] that the other agency is providing you with free education, [so] all you need is supervisor approval.

Joe Polastre created the original proof-of-concept at the State Department, Gwynne Kostin set the agenda at the General Services Administration (GSA) and staffed the program, and Lisa Nelson did all of the super-important people stuff. Joe handed off to me in 2014, [and] I set the technical strategy, led the technical team—sometimes only me—that ultimately led to the platform being adopted across 60 federal agencies [by the time] I left in 2017. It’s now in wider use through the transition to the Office of Personnel Management and is being piloted in the state of California. I also wrote some of the code, but that wasn’t my biggest contribution.

Open Opportunities was one of the most successful digital transformation initiatives of that era, where the software played a small, yet critical, role, empowering a lot of federal employees to amplify their impact by developing skills and working together on mission-driven priorities across government.

What have been the biggest changes you’ve experienced in your career as a frontend and web developer? What do you think the future holds?

The separation of visual design, layout, and user interface elements from information conveyed through the interface is not only essential for accessibility, [but it] also makes it easier for people to export to different formats and for software developers to transition to new technologies. There are parts of this in the JAM stack (JavaScript, APIs, and Markup). It could all still get entangled again, but the separation of each discipline accelerates development, makes it easier for teams with different skills to work together, as well as provides a forward path for the so-called “content” if the software dies.

Video took a step backwards in the past decade with what I believe was a bit of an overinvestment in “video on demand” over HTTP. I understand why, but it was easier to create low-latency, highly-interactive experiences with video in 2002 than it is today. It’s amazing to me that RTMP is still widely used [for] most services that support live video. I think we’ve turned a corner with browser capabilities and WebAssembly that can deliver high-performance client-side code in the browser. Maybe we can finally have interactive video in the browser that is safe, works well, and is based on open standards.

This chimes with what you were saying earlier about how software needs to be open and available in order to truly be alive.

Software is everywhere. That’s not an altogether good thing. Farmers now buy tractors which they are not permitted by law to repair themselves. We’re creating a very fragile society by allowing companies to sell machines that just stop working. Each of us needs to take responsibility for the software we write. When I worked on open source in government, I felt it opened up a new path for participatory democracy. With our decisions about what we restrict and what we enable, we can encode freedom. Today and in the future, I’d like to see more software that allows people to create and contribute to shared knowledge and tools.