by Jim Dwyer
“We finally have the building blocks to build out this social web,” Tantek said. “It really feels like we have been empowered with all that has gone on with you guys.”
“We’re definitely pushing as hard as we can,” Max promised, but suggestions for fixes, “patches” to the codes submitted by volunteers, were coming in so fast they were inundated. It was an embarrassment of riches.
“We had been the four dudes just working it out, but at this point, we’re totally overwhelmed,” Max said.
To Tantek, the stakes went beyond the logistics of one group of kids struggling with attention and one another. Using the web was like walking on soft grass, leaving traces of every step. For the indie web troops, the stakes were control over what footprint you leave in the world.
Tantek had some advice on getting clarity on what needed to be done first: eat the dog food, reiterating what he had said in Portland.
“Running it on your own site will prioritize that stuff so fast,” Tantek said. “The best programming comes when an individual has an itch, and they do whatever it takes to scratch it.”
Max was a heavy, heavy user of Facebook, but not of Diaspora. It was a running joke in the group that while building Diaspora, he was still checking his Facebook page every two minutes.
But no dog food, as yet. In fact, he had not yet taken Diaspora out for a real-life drive. “I’m interested to start using it,” Max said. “And Dan is, especially. Dan is the one who doesn’t have a Facebook account. He just feels totally isolated from people. He’s desperate.”
“And he’s resisted getting an account?” Tantek said, impressed.
No, actually. Dan had had one, but he vaporized it.
—
Dan spent a weekend in Providence, Rhode Island, about two hundred miles north of New York; he was invited to lead a panel at a conference called “A Better World by Design.” He and the audience, which included many students from Brown University and the Rhode Island School of Design, talked about online relationships and what they had to do with real life.
As the session wound down, a young man raised his hand.
“Do you have a business plan?” he asked.
“We don’t have a business plan,” Dan said. “We just want to make software.”
—
Break over, they returned to San Francisco for an October of humming days and nights.
One evening, the scent of hot pizza, still steaming from the oven, set off a small-scale stampede toward the Pivotal kitchen. Floating through the windows, a raucous din rose from Market Street. It was a time of day when both the streets of downtown San Francisco and the offices of Pivotal software were normally empty or close to it. But on that Thursday night near the end of October, the San Francisco Giants were playing the second game of the 2010 World Series. It was far more fun to take in the game with company, even strangers, in bars, or, if in a car, to lean on the horn.
At Pivotal, the meeting area was filled with people paying no attention to the baseball game at all, but who wanted to help Diaspora. Among them was a professional software developer named Sarah Mei, who had convened the gathering.
Even as clouds of venomous criticism of Diaspora blew across the Internet, an ant colony of people had started working on it from all over the world. Most of the time, they could be seen only in trails of terse comments left on GitHub, the repository, or in Internet chat rooms. Sarah Mei had tightened the code immediately after its release in September, speeding it up. Ilya and she struck up an online conversation.
“I actually work at Pivotal,” she said.
“Are you in San Francisco?” Ilya asked.
She was.
Ilya digested that for a moment. She had been down the hall the whole summer, not on the other side of the world.
“We should get lunch,” he said.
She had attended the presentation they’d given in the summer, and had taken a hard look at what they had done. First, she was impressed with how undefensive they were about criticism. They did not seem capable of getting their backs up: a decidedly unmacho and—she thought—smart way of handling things.
Second, the code was in much better shape than she expected. So she began to pitch in, quietly, without introducing her physical self to them. Though she was still working for Pivotal, her regular contributions to Diaspora made Max, in particular, happy. Hers was an experienced and authoritative voice, precisely what they needed. He wished he could hire her, but at Pivotal’s rate of fifteen hundred dollars per day, she would burn up their funds in a few months.
In October, Sarah announced plans to hold a Meetup, “Contributing to Open Source 101 with Diaspora.” Meetups were just the kinds of live, in-person gatherings that were never so easy before people were online all the time, supposedly losing their ability for normal human discourse. This one was completely booked within a few days. Organized through a social website called, not surprisingly, Meetup.com, which had been created in 2002 by Scott Heiferman, Meetups propelled the candidacy of Howard Dean for the Democratic presidential nomination in 2004. They had many purposes beyond politics. People enrolled to be alerted about get-togethers on subjects they were interested in—weight loss, politics, fishing, knitting, hiking, technology, and thousands of others. It was an excellent tool for Diaspora; groups interested in it held Meetups in about a dozen cities around the world. Their friend from Stanford, Yosem Companys, had been a big advocate of getting them going. The one in San Francisco would prove to be among the most active.
Rafi was at the podium.
What was their next goal for release?
“We have a goal to release something—an alpha, really—before Thanksgiving,” Rafi said.
Rafi explained that they had a website, bugs.diaspora.com, where people could let them know what needed to be fixed. With slightly more than a month of responses from all over the world, though, he asked his audience not to be too quick to sound the alarm.
“When you find a bug and you want to submit it to a bug tracker, the worst thing to do is to immediately put down ‘New Bug,’” he said, because they were getting multiple reports on the same problems. “A lot of time is going to be sunk into figuring out, no, we don’t have 220 bugs, we actually only have 50 bugs, because really there’s one incredibly aggravating bug that 30 people have submitted.”
Then a man sitting near the front spoke up. His name was Sascha Faun Winter, a design and software consultant, and he did not pussyfoot around the main public critique of Diaspora.
“Have you guys addressed publicly, or otherwise, the security concerns?” he asked.
A small smile seemed to live perpetually on Rafi’s face, a look of bemusement. It did not fade even as he took on this most basic, and possibly fatal, challenge to Diaspora.
“When we first did our release, there were a couple—really, just one—blog posts about security holes,” Rafi began, giving background to the rest of the crowd in the event that they had not seen the criticisms before he directly answered Winter’s question.
“Basically, what we did with that was we fixed most of them—I mean, we already have security warnings on the site and put a bunch more on—and then put the rest of the holes on our bug tracker and have been fixing them as we went along.”
The four dudes shared a striking quality of poise in public, even when faced with criticism that felt excessive. Rafi’s face was the picture of calm. It was part of why Sarah Mei was willing to take time out of her life for their project.
Rafi reached into the core of the critique: “The assertion that it would be impossible for us to build a secure application is just—sort of, sort of bizarre. Maybe it would have been impossible for us if we were locked into a room, trying to get security skills out of textbooks or something. But we’re not. We’ve had a lot of help from security experts. We’ve had a lot of help from experienced rails developer
s.”
Another member of the audience broke in to scoff at the premise of the critique.
“Is it possible to create a totally secure app? Nobody’s built it. That’s an unreasonable expectation. What should be reasonable is people bring up security problems that the community can work on. That’s how UNIX got built.”
Before the group started digging into the code, there was one more question about why they were going ahead with an alpha release in a few weeks. Was the goal, someone asked, that people would use it?
“The goal is that people would test it,” Rafi said.
He paused for a moment, looking over his audience, the small smile never leaving his lips.
“Nontechnical people.”
—
After five months in the boarding flophouse, Ilya thought that just about any decent, sociable place would look good. As he toured the three-story house at 757 Treat Avenue in the Mission District at the end of October, every floor was buzzing with young people. They were cooking on grills, playing pianos, talking earnestly. His tour guide was Gardner Bickford, a tall, blond, bearded man who was the nominal captain of the apartment on the third floor, and possessed of the classic manner of the laid-back Californian. Gardner had posted a notice on Craigslist, but he wasn’t sure what to expect; he had just spent two years in New Zealand and was unfamiliar with the savage competition for rentals in San Francisco that followed the bloom of young tech people with money in their pockets, desperate for a place to stay. The first two people answering the ad arrived at the same time.
A little while later, Ilya arrived, and Gardner showed him around. They climbed back up to the third floor.
“What do you think?” Gardner asked.
Ilya didn’t hesitate. It looked like paradise. It even had a name: the Hive.
“I’ll take it,” he said.
A few years earlier, that same building had toppled into one of the holes in the urban world created by crack cocaine. Then it was taken over by a kind of digital collective, Couchsurfing.com, a place for young people on the move to swap digs. Actual glass windows replaced the plywood boards, the alleyways were swept clean of the glass vials that had crunched beneath every footstep, and the place was restored not only to a state of good repair but to one of ample possibility. It shared a sprawling concrete backyard with the adjoining house. There was even a permanent tent/marquee set up in part of the yard, an outdoorish retreat for the damp San Francisco nights. The Couchsurfing people lived and worked in it, including Cory Fenton, the computer programmer who had dreamed up the project. Then Gardner, who had done volunteer coding for the site, was moving back to California, and the couch surfers needed someone to take over the lease. He stepped in. Not long after, so did Ilya. At last, he lived in a place with people who interested him. It was full of promise.
That week, Ilya wrote a post for the project’s official blog. He listed the improvements they had made in the six weeks since the code had been made public. It was easier to send messages out, to post public messages not only to the Diaspora site but also to Facebook and Twitter, and the whole “getting started” experience had been made friendlier.
“Our basic feature set is almost done,” he wrote. “Once that is stable, we’ll set up an alpha server so that anyone, not just developers, can try Diaspora and help us improve it. We’re shooting to do this before Thanksgiving.”
They were off to the races—or so it seemed.
One evening, Ilya walked out with Mike Sofaer, who was on his way to train in a gym two doors away. He stood in the gym lobby and poured out his heart.
Ilya was overwhelmed by the work. He was no good. He did not have the coding skills the others had. Mike was about the only older person he knew well enough in San Francisco to trust.
“You don’t have to be the best programmer in the organization,” Mike said. “Not everyone can be the best programmer.”
Ilya fretted that he could not do the high-level architectural design as well as Rafi.
“That doesn’t matter,” Mike said. “It’s not important that you don’t do it as well as him.”
What mattered, Mike said, was his drive and enthusiasm, things he communicated better than anyone. His spirit. “You’re the heart and soul of it,” he said.
“Thanks, man,” Ilya said.
Mike didn’t think he had persuaded him.
CHAPTER THIRTEEN
The days and nights before Thanksgiving 2010 blurred, an all-out sprint. They had kept the first promise, to release code for other developers by the end of the summer. Facing a deadline for its debut outside that rarefied world, they converted the scoffing nastiness of many critics into high-octane fuel, and stayed up all hours chasing flaws. They would show people, not tell them. That meant paying attention to important bugs that were coming their way from all points on the globe. The nature of the problems could be trivial typographical errors, or misstated function names; they could be errors in logic or security. As the number of people using a piece of software grows, more bugs are found. Virtually all software—whether apps for smartphones or the central operating systems, like Windows, Apple’s Mountain Lion, Linux, and its variants—is revised soon after its release, with updated versions automatically sent to users. Although the companies extensively test their software before launching it, only the experiences of millions of people, under all combinations of circumstances, are adequate to reveal significant problems. That point is taken further under a famous maxim of open-source development, called Linus’s Law in honor of Linus Torvalds, the mastermind of Linux.
It holds: “Given enough eyeballs, all bugs are shallow.”
That is, crowdsourcing would not only identify the problem but help fix it. That was often more article of faith than established fact; the vast majority of open-source projects had only one or two people working on them, even though their code was publicly available. But there were a few historic exceptions, most prominently Linux and Firefox. As it happened, the volume of contributors to Diaspora was extraordinary. And they had plenty of things to fix.
The plan was to send out the first batch of one thousand invitations on the Tuesday before Thanksgiving, giving them Wednesday to get out of town and on their various ways home. Each of those invited would have five invitations to distribute. Most recipients were not software developers but civilians inspired by their cause. Real people.
He wrote: “We are proud of where Diaspora is right now. In less than five months, we’ve gone from nothing to a great starting point from which the community can keep working. We’ve spent a lot of time thinking about how people can share in a private way, and still do all the things people love to do on social networks.”
Then he broke news about a feature of Diaspora that would distinguish it from existing social networks: “Diaspora lets you create ‘aspects,’ which are personal lists that let you group people according to the roles they play in your life. We think that aspects are a simple, straightforward, lightweight way to make it really clear who is receiving your posts and who you are receiving posts from. It isn’t perfect, but the best way to improve is to get it into your hands and listen closely to your response.”
This was a striking, if obvious, innovation. Up to that point, anything posted on Facebook, be it a picture or a comment on a news story, was visible to all of a user’s “friends.” It was possible to block the general public from seeing it. But Facebook and most social networking sites did not allow users to exercise anything like the kinds of granular control real people used in their everyday interactions—the instantaneous, natural grading of communication with different individuals, the changes in inflection, in tone and topic, when speaking with members of a softball team or friends from work, high school classmates or close relatives. Diaspora introduced a simple way to allow the world to be divided beyond the crude boundary of “friends” and “nonfriends.” As people were brought int
o the network, the user could tag them, essentially sorting them into different buckets. A post could be distributed to any or all of those buckets. That meant a message meant for drinking buddies would not end up in the news feed sent to Mom and Dad. The subgroups in the buckets were called aspects, as in the different aspects of one’s life.
The creation of aspects was a necessary contribution, and much in keeping with their goal of not just improving geeky stuff like privacy and data ownership but also making communication richer and deeper.
The post, signed by all four, promised that they had fixed the security bugs that had gotten so much attention in September, and that they were going to continue to tighten the hatches. They also were going to put together better documentation, clean up the code, and make it easier for all the people hosting Diaspora pods to upgrade each improvement. The post concluded: “Our work is nowhere close to done. To us, that is the best part. There are always more things to improve, more tricks to learn, and more awesome features to add. See you on Diaspora.”
The tech blogs exploded with the news, and tweets began flying. The announcement had been a bit premature: before they could distribute it, they needed signed agreements from the people who had contributed code, agreeing to release their copyrights to Diaspora Inc. This took e-mail hand-holding because the license they were using was thought to be a mess. Finally, though, the “pre-alpha” version was released late that evening.
The response was vigorous and lively, the hissing and sneers drowned out. They had kept their promises: to release a version to developers in September, and now, two months later, an alpha for the public. Diaspora was not vaporware. It took a little while for the public to absorb the reality that Diaspora was not a website; it was a piece of software that ran a website. Someone had to host the website, provide the servers, maintain them. The supporters who received the early invitations could sign up on a website that the guys themselves were hosting, JoinDiaspora.com. But the guys had no intention or desire for this to become the only place where Diaspora existed. Anyone could host a site. The software was free. While it was raw and buggy, precisely what would have been expected at that stage, work on it would continue. All fixes would be shared with anyone hosting an instance of Diaspora. The very nature of their plan was that they would not be in charge of everyone’s data: they were drafting the DNA of a distributed social network, not growing and feeding the creature.