Harvard Institute of Technology
Source: C. Ramsey Fahs

Disparities between the computer science communities at Harvard and MIT are probably due in large part to the way each was programmed. About 100 years ago, Harvard was programmed, more or less, to neglect Engineering in general, and Computer Science in particular.


Brian “Foundation” Shimanuki wins another game.    After shaking hands with his opponent, Foundation walks back to his seat with a grin. An enthusiastic crowd cheers on their clear favorite.
It is 9 p.m. in Cambridge. It the last Saturday night in January, and this is the Battlecode finals.
In the game of Battlecode, a red team and a blue team face off.    Each team aims to kill the other’s archons—special units that can create other units—while protecting its own archons from attack. Teams employ various basic strategies (defend and gather the resources scattered around the map, attack early and often in the hopes of a quick victory, etc.) and endless minor permutations of those strategies. It’s a complicated, slightlier nerdier sort of chess.





At MIT’s Battlecode competition, founded in 2001, teams create computer programs which control every decision made during a round, with the goal of defeating the other team’s “archons”.
Cynthia Guo
Unlike chess, though, Battlecode, founded in 2001 and run by MIT students, is not actually played by humans. At least, not directly. Instead,    two computer programs control every decision made during gameplay.
Teams use the month of prep to strategize, but they also spend it translating and coding that strategy into Java. It’s not the two teams that are competing:    It’s their respective programs. By the finals, results are essentially decided. If there’s a problem in your strategy now, it’s too late to change it.
The big twist this year is zombies. Zombies attack any unit, red or blue, within their line of sight, and they get stronger as the game progresses. Foundation’s whole strategy centers around zombies. Instead of creating his own army, he only creates scouts, units that are very fast, but unable to attack. The scouts find zombies and lead them to the enemy.
Foundation ultimately places fourth, but the crowd adores him, and his competitors fear him. On Foundation’s “team wall” on the Battlecode website, an adversary and admirer writes, “You twisted, cruel bastard. 10/10 strategy.”
Though MIT students organize the event, competitors from all over the world can submit code. The final tournament, an hours-long, 16-team affair, features teams from Germany to nearby Harvard.
The Harvard team, “what_thesis,” has far fewer opportunities for such CS-centered gaming on their own campus.
Indeed, when it comes to computer science in Cambridge, there’s little doubt that MIT has the more established culture and curriculum. It boasts more students, more professors, more interest from recruiters, and much higher rankings. Differences between the two schools don’t stop there. Where Harvard’s campus is picturesque, MIT’s is stark. Where Harvard students have “concentrations,” MIT undergrads major in a number—6 for Computer Science, 2 for Mechanical Engineering.
Much like in Battlecode, though, disparities between the computer science communities at Harvard and MIT are probably due in large part to the way each was programmed. About 100 years ago, Harvard was programmed, more or less, to neglect Engineering in general, and Computer Science in particular. To understand that Harvard’s computer science department seems comparatively underdeveloped, look to the man who coded Harvard: Harvard’s 21st president, Charles William Eliot.
The New Education
Maria Stoica ’17    checks her email in the Eliot House dining hall. Though she seems at ease, Stoica is not the sort of student Eliot, Harvard Class of 1853, would have imagined at Harvard. While he certainly opposed allowing women at Harvard, Eliot’s real problem with Stoica would have been her studies. Stoica is a computer scientist, an engineer. Harvard is a liberal arts college. In Eliot’s mind, those things shouldn’t mix.
“The practical spirit and the literary or scholastic spirit are both good, but they are incompatible,” wrote Eliot in an 1869 piece in The Atlantic dubbed “The New Education.” “If commingled, they are spoiled.”
This article, many believe, landed Eliot the Harvard presidency. He penned it as a chemistry professor at the Massachusetts Institute of Technology in Feb. 1869. Months later Eliot assumed the Harvard presidency. Stoica also made the move from MIT to Harvard, transferring between her sophomore and junior years to find a “well-rounded ” culture that she felt MIT lacked.    She wanted the kind of “co-mingling” of disciplines that Eliot so excoriated in 1869.





Maria Stoica ‘17, who lives in Eliot house, transferred to Harvard from MIT after her sophomore year to find a more “well-rounded” culture that she felt MIT lacked.
Cynthia Guo
Over the course of his presidency, Eliot tried many times to absorb MIT into the Harvard umbrella, coming closest to success on his last attempt in 1905 before MIT alumni and faculty successfully rallied against the merger.
Perhaps frustrated by his last failure, Eliot absorbed Harvard’s formerly independent Lawrence Scientific School into the College in 1906. The move marked    a complete reversal from Eliot’s position in “The New Education.” In that essay, he advised against putting “scientific schools” in the same organizational wing as the more established fields (the classics, philosophy, etc.) A scientific school within a college, Eliot wrote in 1869, “is the story of the ugly duckling.”
The younger Eliot was right. Harry R. Lewis ’68, professor of Computer Science and former interim Dean of Harvard School of Engineering and Applied Sciences, blames the Lawrence Scientific School-College merger for “effectively put[ting] engineering and applied science in a backwater” for close to a century. With some notable exceptions, Harvard engineering remained an ugly duckling.



Harry R. Lewis ’68, professor of Computer Science and former interim Dean of Harvard School of Engineering and Applied Sciences, had proposed to have a computer science concentration in around 1978.
Cynthia Guo
In the mid-70’s when MIT (1974), Stanford (1965), and Berkeley (1968) already had Computer Science departments, Harvard remained opposed. In around 1978, Lewis, then a non-tenured assistant professor, proposed a computer science degree at an Engineering faculty meeting; computer science was part of the Applied Mathematics department at that time. According to Lewis’s blog “Bits and Pieces,” Bernard Budiansky, an applied mathematician, opposed the idea, asking, “We’ve never had a major in automotive science, why would we have one in computer science?”
Budiansky’s gripe seems straight out of Eliot’s writings. Eliot was fine with teaching most subjects, but “they should be taught at the university on a higher plane than elsewhere.” Computer science, in Budiansky’s mind, was a practical and technical skill. It wasn’t until 1984 that Harvard awarded its first undergraduate degree in Computer Science, and by that time other programs had almost a decade’s head start.    By and large, the early adopters (the MITs, the Stanfords, the Berkeleys) are the top programs today.
In colleges across the country, the debate continues over what computer science actually is. A mélange of poorly defined terms (coding, hacking, computer programming, computer engineering, etc.) are routinely used to describe the field and a bevy of online services (Codeacademy, Code.org, Launch Academy, etc.) advertise a quicker, cheaper route to a job than a four-year degree in computer science.



The Digi Comp II in MIT’s Stata Center, which houses the Computer Science Department, is a toy computer which performs basic calculations using balls rolling down an inclined plane.
Cynthia Guo
A Wall Street Journal op-ed published last August titled “Why I’m Not Looking to Hire Computer Science Majors” blamed computer science departments for not adequately preparing their students for the industry. The op-ed criticized the lack of iOS and Android development classes in Ivy League universities like Yale and Harvard. This is a common enough complaint at Harvard, and even at MIT, where students believe the department favors the theoretical over the applied. For Tomas A. Reimers ’17, though, studying computer science is like studying Spanish literary theory: if you’re going to learn it at all, you have to learn it in Spanish.
“There’s… this, ‘You need to learn Spanish’ portion, and that’s what programming is,” says Reimers. “Then there’s proper CS, which is how you think about data structures and algorithms and things like machine learning and visualization.”
Regardless of how you define it, computer science has become immensely popular. Harvard’s Computer Science department climbed from 69    concentrators in the 2007-2008 school year to 348 this year, with an inflection point coming, a chuckling Lewis points out, right around the release of the Facebook-inspired “Social Network.”
A $50 million dollar gift to Harvard Computer Science from former Microsoft CEO Steve A. Ballmer ’77 (with the Ballmer-esque edict to put Harvard “on that list” of top computer science schools alongside MIT, Stanford, and Carnegie Mellon) and a record-breaking, school-renaming $400 million from Business School alumnus and hedge fund manager John A. Paulson have given Harvard the capacity to grow computer science like never before. Throw in an historic $1 billion move to Allston, which Lewis has called “the Promised Land,” and the CS department is looking at a future of sustained growth into uncharted territory.
Harvard’s neighbor down Mass. Ave., though, has been running massive computer science operations for a while. Harvard CS won’t reach the size of MIT’s department any time soon.
Its growth, however, has made the University question the way it has operated since Eliot’s tenure. What is the ideal balance between extracurriculars and scholastic pursuit? What is the proper amount of corporate involvement in a university education? What is the best way to introduce students to a subject? MIT has (at least in part) answered many of these questions, and Harvard must, too, as enrollment and investment in computer science reach a fever pitch.

Here for the Classes?
At MIT, students mostly engage with the computer sciences through their academics. Harvard CS, on the other hand, is as much about what happens outside the classroom as it is about what happens in it, with Harvard students choosing to focus more on the extracurricular side of their CS studies.
None of the real rockstars of Harvard CS, after all, ever graduated from the school. Both Bill Gates and Mark Zuckerberg quit early to commit to their respective start-ups full time.
Even those who don’t go so far as to drop out, though, look for ways besides academia to pursue computer science. Between working as a teaching fellow, joining an organization like the Harvard Computer Society, or working for fun on side-projects or start-ups, Reimers says computer science “allows you to focus all of your energy on it.”
“If you want to dump your life into computer science, [the subject] very much encourages that,” says Reimers. “In fact, it rewards it.”
But then, Harvard students are notoriously committed to life outside of the classroom. The proportion of students who self-reported spending more than 10 hours a week on extracurriculars rose from 27 to 43 percent between 2010 and 2013, according to data from the Senior Survey.
According to Divya Shanmugam, an MIT junior in Computer Science, life is the classroom. “I would say that classes take up a majority of your time, and then if you manage everything well you can do extracurriculars,” says Shanmugam. “I’m starting to quit my extracurriculars and say, ‘I am here just for the classes.’”
Shanmugam was drawn to MIT as a high school senior in part because of the “starkness” of the campus and the “no frills education” that it seemed to promise. This ascetic ideal of pure academics is foreign to MIT Computer Science professor David R. Karger ’89, who took music classes and sang in a choir during his undergraduate years at Harvard.   
“[Some] people take nine courses in a semester and ace them all ... It’s very hard in that environment to hold back and say, ‘I want to have time,’”    says Karger.    “I can’t imagine having time to do that with the kind of course loads that MIT students face.”
MIT also boasts a conventionally stronger department, mostly due to its size. When it comes to computer science departments, size matters. Harvard CS, according to new Harvard hire and former MIT professor Madhu Sudan, has always had a reputation for strong individual professors (including two winners of the Turing Award—a sort of Nobel Prize for computing) but a weak overall department. It lacked “critical mass.”