When Charlie Croom ’12 entered the world of programming, he was six years old and had just learned to read. When he switched on his mom’s computer, he was met with a blank screen and a blinking cursor. The machine ran on floppy disks and a dial-up modem, and only understood commands typed out in code. Croom typed until he figured out the classic line that starts a programmer’s career:

PRINT “Hello World”

One day this spring, Charlie Croom was sitting in Yale’s computer science class called Design and Analysis of Algorithms, and he wasn’t listening. He had a laptop in front of him with a screen that glared white, and he was typing out thin black lines of code. Everyone around him stood up — the class was over — so he closed his laptop, stood up, sat down in his next class, and pushed the screen upright again.

Fourteen hours a day proceeded like this: Croom face to face with ordered black trails of thought. Croom and his suitemate, Jared Shenson ’12, were coding Yale Bluebook, their student-run redesign of Yale’s online course selection service (OCS). Their goal was to get Bluebook ready for the first Yale College Council App Challenge, whose deadline was looming in two weeks.

Shenson and Croom kept tweaking Bluebook long after it was one of four finalists in May — they were still coding in August when thousands of students logged on to their site for class searches instead of OCS. They kept polishing the menus and centering text while Yale’s technology services considered buying the site and while the YCC pitched the program to University President Richard Levin, who was enthusiastic.

Shenson and Croom are not the only ones with recent success. Booksaver and Books@Yale both streamline coursebook purchasing, and Booksaver has expanded to over 400 schools and earned $100,000 from private investments. Roammeo won the App Challenge for streamlining campus event listings; SubletMeYale gives students an alternative to Craigslist; and the half-joking One Button Wenzel site led to so many orders for Wenzels that Alpha Delta Restaurant had to bring in an extra employee through the first weekend the site went live.

Shenson, Croom, and the other finalists of the YCC App challenge are at the center of a startup culture that has expanded from a small number of isolated entrepreneurs to dozens of excited newcomers. The startup culture has the potential to re-imagine every online aspect of Yale student life. It has led some computer science majors to forgo their problem sets and re-think applying for traditional software firms, instead gambling on their own startup ventures. Meanwhile, if Yale buys Bluebook, it will have invested in software that was built during time spent ignoring the University’s own classes.

Two dozen white screens are glowing inside a modest lecture room on the highest floor of William L. Harkness Hall, where Bay Gross ’13 and Will Gaybrick LAW ’12 are teaching. The two have led the creation of HackYale, a organization that teaches the basics of designing web applications and received nearly 500 applications for 50 slots.

“We are taking a very different approach than most computer science departments,” Gaybrick says, staring out at the faces behind the screens. “We’re focusing on the practical. Learn by doing and learn by immersion, not by listening to us.”

The most popular programming languages once looked like streams of backward-leaning slashes and semicolons, but coding for the web has started to sound more and more like English. In the ’90s, languages cost money to use and required slogging through dense books; today the succinct and friendly Ruby on Rails, for example, can be learned through an abundance of website walkthroughs. Gaybrick and Gross have put their lecture slides online for free, along with all of the course’s resources and homework. Both of them hope their course will inspire a community of coders that did not exist six months ago.

“[Yale] could be an absolute hotbed for startup companies,” Gaybrick later said. “The scene just needs to be jump-started.”

Across campus on Whitney Avenue, two dozen students meet to discuss their startup ideas at the Yale Entrepreneurial Institute, whose main space has soaring rows of shiny bright wood tables atop blue carpet and fluorescent lights. There are very few chairs. A whiteboard against one wall is littered with “Forbidden Corporate Buzzwords” like “synergy” and “leverage,” and a poster of Thomas Edison in the corner reads, “SPEED VICTORY / Let’s have your ideas.”

The students who have gathered in the Institute’s meeting space sit mostly by themselves, working on their computer science problem sets or surfing websites that introduce programming. Two students confirm that they are working on startups, but they decline to talk about their ideas. A few physics graduate students sit toward the back, softly discussing something about a “Facebook replacement.”

Before the sudden rush of interest in programming websites, the traditional way to learn about computers was through the computer science major. Yale’s building for computer science is Arthur K. Watson Hall, but its main computer lab has been dubbed “the Zoo,” where the computers are named after animals.

Computer science classes center around pulling up a box on your screen, inputting code to match the theory and getting numbers to spit back at you. Classes cover subjects like graph theory and data structures, and the more advanced problem sets are often elaborate programs that each student sends to a professor, who runs his own program that tries to find holes in what each student has created. Depending on how well you’ve coded, very few or no things will make your program fail — and your grade is determined as quickly as the holes are found.

One of the major’s rights of passage is the LZW problem set in CS 323, which requires each student to compress large amounts of data in order to consume less memory. Gross spent 25 hours over one week on the problem set, and three computer science majors confirmed that some can spend 50 hours over two weeks.

The classes can lead to a senior project that borders on abstract math. Cameron and Chris Musco ’12, identical twins who are both computer science majors, are pursuing projects that involve finding patterns across huge data sets. The process is something like finding a best-fit line across three dimensions.

But outside of class, Cameron and Chris taught themselves something far more concrete: website design, which they used to code One Button Wenzel. “Most startups address everyday problems,” Cameron says. “Theoretical stuff might be less useful in the long run — your algorithm might never be used. But there’s a certain intellectual gratification from the math.”

Chris adds, “In startups, your final product is appealing to the everyman. In academia, when it’s used, it will be hidden. It’s satisfying to prove something that someone else couldn’t prove, or make a connection that someone didn’t make.”

CS 323 is taught by Stanley Eisenstat, the department’s director of undergraduate studies. Eisenstat came to Yale to teach computer science in 1971, when computers had just started shrinking from the size of rooms. He has a gray beard, physicist’s glasses, and wears white New Balances. According to Eisenstat, the Computer Science Department teaches languages for the theory behind them, not for their usefulness in modern programming. Choosing one language, he says, can be self-defeating once it gets old in a few years, and serious websites run on a host of languages. So Yale’s Intro to Programming has been teaching the desktop language Java since the mid-90s, which Eisenstat says is best for new students because of its “sandbox” nature — it allows students to experiment without accidentally rendering their programs useless.

But other universities have approached web languages differently. Harvard’s Computer Science 50 introduces website languages like PHP and JavaScript in addition to the traditional desktop language C. All of the course’s problem sets are online for free and the lectures are on YouTube and iTunes. Even though students reported about 11 hours spent on homework each week, the class has just over 600 students, making it Harvard’s second-largest undergraduate enrollment this fall. It lost only to an intro course on microeconomics. The course ends with a “CS50 Fair,” where students present their final projects to the college at large. Gaybrick and Gross said HackYale was inspired by the course’s open-source concept.

“We want students to realize that they’ve not just learned C; they’ve learned how to program,” the Harvard course’s senior lecturer, David Malan, said in an email.

And near Silicon Valley, Stanford has offered a course on designing iPhone applications since 2009, which has been taught by Apple employees and whose free lectures, according to Stanford’s website, have received over one million iTunes downloads. Malan plans to introduce a similar course at Harvard this spring.

The coursebook-buying site Books@Yale’s founder, Sean Haufler ’13, is glad that Yale’s major teaches the more abstract concepts, and says they have enabled him to pick up the simpler web languages in a matter of days instead of weeks. But Haufler wishes the department would consider an intro class in basic web languages. “[CS50] is the perfect example of what should be done at Yale, and what hundreds of students would like at Yale,” he said. But Eisenstat says that interest in programming has had its high and low points. The number of computer science majors has risen sharply, from 15 graduating in 2010 to a projected 30 in 2013, but there were about 40 graduating in 2004, and before that, the 2000-era dot-com crash brought those numbers into the tens. “In the 40 years I’ve been in the field, it’s ebbed and flowed,” he said. “I expect it to continue to ebb and flow.”

Geoffrey Litt ’14, who helped design Bluebooker — a four-year course planning app that was an App Challenge finalist — wishes the department would consider the Internet’s role more strongly, adding that a student can complete the major without knowing the basic web language HTML. “The fundamental paradigm of how things work on the web is different,” he says. “Of course Ruby’s going to disappear … but the paradigm of how you program for the web has been established.”

Web languages are becoming even easier, and at a faster rate. They are easier to learn than anything spoken by humans — once a programmer has figured out the basics, languages ask for instructions in very similar ways, and writing new lines will often consist of tweaking ready-made snippets from a quick Google search.

“Ninety percent of programming is just Googling,” Croom says. He shakes his head at his own analogy: “You make Shakespeare by looking up every single word.”

As for whether Yale’s Intro to Programming will ever start teaching new languages, Eisenstat says the class may switch to Python, a web engineering language that reached its peak of popularity about six years ago and has since been largely supplanted by Ruby on Rails.

“Computer Science departments have historically been about developing programmers for research applications … the next operating system, some high end medical device,” Jared says. “As web apps become increasingly popular and relevant, that’s where people feel they can have the most tangible impact. Computer science departments have not accepted that fact yet.”

Jazear Brooks ’12, a computer science major, has a lucrative job offer from Microsoft. But last year he was chasing startup dreams on his own.

He took last year off, got a job as an SAT tutor in New Haven, and spent the rest of his time working with two New York-based entrepreneurs he met at a Manhattan startup conference. The three of them devised a website called Kinecticle that would coordinate Facebook users’ workout schedules. A venture capitalist was interested in funding their work in the six-figure range, but Brooks realized he would need to teach himself Ruby on Rails if the project was going to grow.

He first tried searching for New Haven-based programmers to help him walk through the idea. He showed up to a meeting for a group of professionals called Newhaven.rb, but the members were hard at work on their own projects.

“I ended up realizing if you want to build something, you have to build it yourself,” Brooks said. “It’s too technical to have someone walk through what I wanted.”

So Brooks taught himself from online resources like

railstutorial.org, but when he tried to integrate the site with Facebook, he found himself caught up in a trial-and-error game that wasn’t leading anywhere. (He added that Facebook has since made it easy to fix what he had been grappling with.) Months passed, and suddenly it was spring. The venture capitalist offered Brooks and the team a six-figure deal if they wanted to start talking about contract agreements, percent ownership, and hiring a bigger development team. But Brooks still hadn’t integrated the site with Facebook, and even once that was fixed, he would have to code for much longer than a year. At about the same time he got a 12-week paid summer internship at Microsoft, a popular career path among computer science majors at Yale.

Brooks chose Microsoft and moved to its headquarters in Seattle, where he worked on the next version of Microsoft Office. The startup team declined the funding and suspended the idea indefinitely.

Brooks now wishes he had started with around $10,000 of investments instead of aiming directly for the serious capital. He says he should have talked to more Yale students about how to manage his year off and tried to move the project to General Assembly, a Manhattan startup center run by Yale alums that has been called “the shiny new hub of New York’s startup scene” by the New York Observer. “Had I [started] a year or two younger I would have had more experience,” Brooks says.

Even though Brooks helped start Yale Hackers, he says the surge of student interest in programming is likely short-term. He compares it to brief crazes in the 2000s that died out almost as quickly as they arrived. “In the long term, computer science will always subsume web development,” he says.

Meanwhile, Sean Haufler ’13, founder of Books@Yale, is entering the startup world on his own. He came to Yale as a varsity swimmer and taught himself how to code with the free online lecture series from Stanford and Harvard. He first found a niche on the web when he noticed that an Italian company selling one of the best swimsuits in the world, Jaked, had no online shipping to US swimmers in time for the 2009 World Championships in Rome. Haufler emailed them for the rights to sell their suits through a website he had just created, and they agreed. He sold 100 suits in three months and noticed some of the credit card orders belonged to famous American swimmers — some of them his swimming


Haufler was browsing Yale’s online course selection early this summer when he realized Yale had catalogued the ISBN data for all the course books into one place. But no one used the data because the site redirects to a page for each book at Barnes and Noble, when students would often rather use Amazon. He built the bare bones of Books@Yale in four weeks on his living room couch in San Francisco when he wasn’t working as an intern for a San Francisco startup. After a total of eight weeks, his site was live in time for Yale’s rush of book-buying. Haufler recruited a friend at Harvard to email final club panlists about his new Harvard version of the site. Through promoter deals, Haufler has since expanded the site to over 60 schools and is adding new schools at a rate of three each day.

Croom had been staring at code for the last three hours, trying to ignore hundreds of people as they filed into a basement auditorium in Manhattan. He was double- and triple-checking code for a startup demo at the New York Tech Meetup, which selects six startup teams each month to show their web apps to venture capitalists and the public. Croom is the lead coder for one of those teams. It was approaching 7 p.m. and he hadn’t eaten since he caught a train from New Haven early that afternoon, but food was a low priority. He and the startup’s CEO, Seth Bannon, had just three minutes to prove what the site could do. If a single drop-down menu failed to expand smoothly, investors would pass them by.

Croom and Bannon’s nonprofit campaigning application is Amicus, which builds phone banks based on Facebook friends. The idea is that campaigners should not have to pour through phonebooks alphabetically: an earlier version of the site carried the slogan “Killing the Cold Call.” The web app can also create door-to-door lists based on similar techniques.

Amicus is almost entirely run by Yale students and graduates. It is coded by Gross and the Musco twins, designed by Shenson, and managed by a Yale alum. Gross and Croom spent the summer working on the site at General Assembly.

The Amicus team has been testing the site with seven nonprofits and campaign groups, including the New Haven Ward 7 aldermanic candidate Doug Hausladen. Despite some short-term investments, the team is looking for funding to support themselves through December 2012.

In the Manhattan auditorium, Croom and Bannon finally climbed up onstage, where New York’s mayor Michael Bloomberg had opened the meeting. As Seth took the microphone, Croom loaded the Amicus homepage onto the enormous projection behind them. He couldn’t stop thinking about the tiny lines of text that weren’t quite centered. Both were anxious: Bannon started speaking extremely quickly, and Croom loaded the pages even faster than Bannon could talk.

Suddenly three minutes had passed, and they were done. There was applause. No one noticed the tiny quirks in the design.

The two were led to an open bar after-party on the second floor of The Foundry just outside Manhattan, where each of the six startups had their own dinner booth for fielding questions from a bustling crowd of investors and techies. The booth was far too big for the two of them. Despite the open bar, Croom did not see any food. They stood in front of the booth and fielded questions from excited nonprofit owners and coders who threw around tech terms like “deviance.” By the time Croom caught his train back to New Haven it was almost midnight, and he still had not eaten.

At a HackYale event two nights later in New

Haven, there are no laptops in sight as Katie Rae speaks to about 40 students in a Timothy Dwight common space. Rae is the managing director for the Boston branch of TechStars, one of the country’s most successful startup incubators. “If you become a big banker, you will make more money than almost anyone will in tech,” Rae tells them. “[But] your life expectancy is much, much lower. When you join a startup, every day is really interesting. It may not be as much money in the beginning, but you’re creating something. Know that your expected joy will be much higher, even when you fail.”

TechStars selects about ten startups at each location from a pool of over 1,000 applicants; the winners are granted $18,000 in funds, support from 100 professional mentors, and sometimes well over $100,000 once the project gets going. In return, the startups give six percent of their profits to TechStars. Nearly all of the companies accepted to TechStars are still active or have been bought. Katie Rae chooses which startups to fund.

Rae later said of building a campus startup culture, “If you get a couple of successes, it starts to snowball really quickly. Yale can do that. But it’s a cultural shift. You have to have a student population that also cares.” She added, “I’m willing to bet that with enough strong students, you could do it without professors.”

Facing the HackYale group, she says, “I need a volunteer who has an idea.”

One student stands up warily in front of everyone. She aims both arms straight out at him: “You have thirty seconds. Pitch your idea.”

Several try their unrehearsed pitches. “The core skill that anyone learns in a startup is listening,” she says. “The best do it really, really well.” She checks her watch. “I have time for one more. Is anyone working on something right now?”

There is a sudden moment of uncomfortable silence. Will Gaybrick, sitting in the front, glances back at Gross, nods, and starts to point an index finger. Gross is the youngest Yale student working on Amicus, providing more than 20 hours a week of coding. When Croom considers which Yale student would take over Bluebook next year, he says, “Probably Bay.”

For a lingering moment, it feels like Gross is going to pitch Amicus, the kind of idea that Katie Rae would consider funding. Gross sits down and faces a room full of interested eyes:

“This isn’t what I’m actually building, but I’d like to build it. I don’t know if any of you have heard of Klout, with a k, like a measure of how cool you are on the Internet? I want to make Klout, but for bros. Instead of how popular you are on Facebook and what your re-tweet potential is, we see how often you say things like ‘keg stand!’ and ‘Natty Light!’ and ‘Brah!’—and then you get a bro score, and you can push it on your Facebook wall.”

“Stop!” Katie Rae yells jokingly.

“It’s gonna be huge!” Gross says.

Gross could have pitched Amicus, but he didn’t. He and all of Yale’s coders, in a culture independent of the university in whose rooms they work, have figured out how to do things on their own.