So Long, and Thanks for All the Theorems
by Derrick Stolee
Since I first learned about graph theory in my intro algorithms class, I have been intensely focused towards learning more about graphs and doing mathematics as a passion and a profession. At the end of this semester, I close the book on that part of my career, leave my academic position, and return to my original plan: developing software.
Why am I leaving?
A very prideful part of myself wants to emphasize that I believe I would be “successful” in academia if I stayed. That is, if success is measured by staying employed, then I have been on-track. In terms of day-to-day happiness and long-term satisfaction, I am not successful.
Due to my own personality, I have developed a very unhealthy emotional connection between my academic work and my feelings of self worth. My way of coping (so far) has been to work harder and longer, and to be worrying about work at all hours. This is not a good way to live.
How did I get here?
It may seem like moving from academia to software is a switch to “Plan B.” Instead, it is more that I am returning to my original plan after a short digression into mathematics.
Software has always been a big focus for me, ever since my dad taught me to program in Visual Basic 3 for Windows 3.11 on an Intel 486 (late 90’s, junior high). I continued programming through high school (with limited mentoring) until I got to undergrad and could really dig in to computer science. I decided to be a double major (CS and Math) because math had been easy in high school and it seemed like a good challenge. I really had no idea what I was doing. (This is a common theme in my life.)
Catching the bug
During my first algorithms course, I learned about graphs and suddenly everything made sense. Graphs were an object that I could easily visualize and understand. I picked up the language of graph theory very quickly by doing deep dives into Wikipedia pages. Suddenly, I started thinking about doing a Ph.D. and doing research on these things. Once I took a full combinatorics and graph theory course sequence, this urge became clear and I focused all of my effort towards a Ph.D. so I could research graph theory.
Once I started graduate school, I suddenly had more access to advanced material and mentors to point me in directions I never thought to go. I was lucky enough to get a research assistantship and a fellowship, so I did not have heavy teaching and could focus on solving problems and writing code.
Entering my third year of grad school, I was pretty full of myself. I had discovered a vocation that I was destined to be a research mathematician and I would change the field with my unique view! Classes were not too hard, I got through qualifying exams, and research was coming together! I thought I was on top of my game and I was unstoppable.
Things get real
Then I was offered an amazing opportunity: in Summer 2010, I went to Illinois to work with the Combinatorics REGS program. There, I worked with graduate students and faculty that completely changed my perspective on where I stood in the research field. Suddenly, I got knocked down a few pegs. I realized that I did not have the international perspective of the research community in combinatorics, I did not compare well to students even younger than me, and I would need to do a lot more in order to be competitive.
And that’s what I did. I did a lot over 2010 and 2011. Late nights. Early mornings.My research expanded and deepened. Research became my first thought and my last thought.
This led to other problems in my life. I gained weight. I became easily irritated, depressed. But it was all temporary! I was going on the job market in fall 2012, and I just needed a big push to get all my papers out before then! Or so I told myself.
During the academic year 2011-2012, I was stressed about the job market, but I made a focused effort to take more time for hobbies, for family, for exercise.
Things were looking up, especially when I got offered a postdoc at Illinois! This was exactly the thing I needed to push my career to the next level. However, this also meant that my wife and I would live in two different states for at least a year. Living apart was harder than either of us imagined, but when we had the opportunity to solve the two-body problem at Iowa State we thought we had found our long-term career location and our forever home.
We have only been at Iowa State for five semesters, so it is hard to say exactly when my attitude towards my work changed, but here are a few big things that I noticed [in order of importance]:
- Due to a large number of collaborations and time-sensitive work duties (such as teaching, organizing workshops, etc.), I very rarely had time to do any software development, which is the part of my job that is most satisfying. [In the past, I would fix this by working late nights and long weekend days.]
- I developed a cynical view of teaching. I care too much about teaching to do it badly, but the time required to teach really well is too high to sacrifice research time. Moreover, when teaching goes badly (or students do poorly on assignments/exams) I have a painful emotional reaction that ruins my week. [I used to think I was a “naturally good” teacher, and perhaps that is somewhat true when I am teaching programming courses but it is not the case with other classes, such as calculus or linear algebra.]
- It has been a while since I truly sat down and learned something that was both new and hard (and wasn’t a theorem I helped prove or background for a paper I was working on). Anyone who pursues a Ph.D. must have some wire crossed in their brain that makes the pursuit of knowledge closer to an addiction than just an interest. [This blog was supposed to be a way to motivate my learning, but it sits far below several other things that aren’t getting done.]
Notice that these items are things that are wrong with me, not the job itself. In fact, I know that working with the Iowa State Discrete Math group is the best possible situation I could ever hope for and I am surrounded by excellent people. That’s part of how I knew that a career change was necessary, not just a location change.
What will I do next?
My family is moving from Iowa to North Carolina, where I will be a software engineer at a big software company (but a small, focused team within that company).
In the meantime, I am focused on teaching my current course and finishing any collaborations that are close to completion. This means I am abandoning a couple projects that were sole-author or are too early to finish in the next month or so.
What about this blog?
This blog will remain public and available as a reference. If I feel like it, I may even write to it again, as math will become a hobby that I can enjoy again.
The blog will also open up to other contributors. If you want to write something about a topic that fits here, then please do so and send it to me. We can even arrange for trusted authors to have editor privileges.
What am I proud of?
During an interview, I was asked to talk about something I was proud of. I gave a terrible, ranting answer about a paper where I thought the resulting theorem was surprising and valuable, but even during the discussion I regretted not picking something else. Here are a few things that I will truly consider my lasting legacy in mathematics:
- I had the pleasure of advising Chris Cox on his master’s thesis. I can take no credit for his research ability; I just pointed him in a direction and he ran farther than I could have. He is now at Carnegie Mellon for his Ph.D. and he will do great things.
- I was a small part of the Rocky Mountain/Great Plains Graduate Research Workshop in Combinatorics which I think is a really cool program that is only starting to build momentum. A lot of interesting projects have come out of the last two workshops, but also a lot of friendships.
- A graduate student once stopped me after a talk to say my presentation was the best one at the conference. He was wrong, but it was a nice thought.
- The blog post Canonical Labelings with Nauty was referenced in a recent research paper: Breaking Symmetries in Graph Coloring Problems with Degree Matrices: the Ramsey Number R(4,3,3)=30, by Codish, Frank, Itzhakov, and Miller. They linked to the post to describe how they encoded edge-colored graphs for their symmetry breaking. Proof exists that at least one person read this blog and benefited in some way.
A Last Bit of Fun
Erdős would say that I’m dying. Mathematically dying.
Here are two alternate titles to this post that I couldn’t keep to myself: