me

A Love Letter to MIT's Blackboards

November 27, 2022

As a new PhD student at Stanford there is much to love. And yet, I tremendously miss MIT's blackboards.

In any job -- including research, most of the work is, inevitably, work. Implementing a new idea. Fighting with the computing cluster. Reading and responding to email. In some jobs, all of the work is work.

But as a PhD student, and a researcher in general, there is a special kind of work that feels very different: the creative dance of figuring out the solution to a problem. The feeling of doing this varies by field, and even by subfield. Sometimes, you disassemble a problem, solve its components, and put them back together. Other times, you just write down your definitions and assumptions and stare at them for a while until an answer emerges. Still other times, you get close to an answer and then patch it until it stops leaking.

I find this work to be both the most stressful and fulfilling part of research. It's stressful because it's important, hard, and unpredictable. You never know how the problem will respond to your probes and attacks. Sometimes, it feels alive. Yet when you conquer the monster you push back the darkness and know you've earned it.

The most important thing is to be able to rapidly and completely focus your attention on some part of the problem, learn something, and repeat. For this, external memory is irreplaceable. Here are its requirements:

1. It should be large, so that you can fit the whole problem and process.
2. It should be very fast to access both the problem and what you've learned.
3. It should be flexible, so that your methods can be too.
4. It should be reasonably quick to add, modify, or erase ideas. However, since you usually generate good ideas more slowly than you consume them, this is not as crucial.

I believe that the lecture halls at MIT, which I used to claim late at night to work on my problem sets or research, met these criteria essentially perfectly. Consider 32-141, a favorite lecture hall of mine for difficult problem sets.

1. Each board is approximately 7 feet by 3 feet, and there are 9 such boards which are easily accessible. Altogether, this represents about 180 square feet of real estate.
2. Due to the fact that the boards slide, one can also keep at least 120 square feet of work visible at all times -- and nothing is faster than having the information right in front of you.
3. Blackboards are, quite literally, a blank slate. One can write, diagram, graph, and code equally well on them. I had also invested in my own set of multicolored Hagoromo chalk, which further increased the space of possibility.
4. Blackboards are reasonably quick to write or erase. I never found this a bottleneck. If anything, I liked that they encouraged deliberate work.

There were three other factors that made these blackboards particularly wonderful. The first is that I like to pace and talk to myself as I think. Large lecture halls are conducive to this, both physically and acoustically, and it's hard to do this in a shared office. The second is that I found the ambiance of these late-night lecture halls to be particularly conducive to focus. MIT's lecture halls breathe discipline, and with the knowledge that my good work would be rewarded with sleep, I rarely procrastinated. Third, and perhaps most controversially, I simply like chalk on slate. Chalk never gives me any trouble, whereas whiteboard markers are finicky and distracting (even if my beautiful AusPen markers lessen the pain.) I like that the feeling of slightly chalky, dry hands represents a physical realization of creative effort. And, by using their tools, I feel a bond with the mathematicians who came before me.

Even MIT's classrooms had tremendous space. Those in building 2 had a remarkable 150 square feet of blackboards, even if somewhat less conveniently placed.

As a PhD student with a shared office, I lack all of this. The largest conference rooms in my building, Gates, have perhaps one-third the available writing space of 32-141 (in deplorable whiteboard form), one-sixth the pacing space, and far worse acoustics. My current substitute is to scatter my desk with printer paper and a few pens. It works alright, but it isn't the same.

To MIT's blackboards: you are not forgotten. I miss you.