Jonas Kersulis     Posts     Research     Teaching     Presentations

Schedule Not Deadline

I want to be the kind of person who

  • doesn’t delay in responding to emails.
  • writes 1000 words each day.
  • never misses a workout.
  • people ask for help.

These are identity statements. No numbers attached. I am not saying I want to be able to deadlift 300 pounds by a certain time. I’m just identifying identities I would like to have. This is a key step in forming new habits. Another key step is to make a schedule.

I am driven by deadlines. The two most productive days of my internship last summer were the day before my first talk and the day before my second talk. I did the majority of the research work for my recent conference paper in the month preceding the deadline. I did a great deal of the writing in the last ten days. This process used to work – it’s easy to get an undergraduate degree like this. Years of conditioning got me stuck in the rut of last-minute, deadline-driven thinking. But I got a high GPA as an undergrad and was accepted to multiple graduate schools, so it must have been a decent strategy, right? It was a shock to wake up one day and realize that the deadline mentality doesn’t work in any meaningful setting beyond undergraduate education.

As an undergrad you follow a clearly-marked path determined by others. As a graduate student you make new paths. As an undergrad you are responsible only for yourself; the only name on each homework assignment is your own. As a graduate student you publish papers with co-authors, who must be included in the research and writing processes. As an undergrad you got to bowl with the gutter guards in. As a grad student there are no guards. Sloppy play leads to gutterballs.

Here’s what I wrote on March 12, with precious little time left to finish my conference paper:

My paper deadline is in three days. My code is broken – I have no results. What happened?

People say failure is alright. I say you at least need to understand it. I can blame my lack of results on a deadline-driven work ethic. My MO leaves no room for hiccups and rewrites. I routinely do more than half the work for a project in the week before its deadline. This is fine for homework –and homework is what I have been conditioned for– but it is a terrible research strategy.

When a paper needs the approval of four co-authors, the writing process should go something like this:

  • 3 months left: Abstract accepted. Begin research work.
  • 2.5 months left: Have a working problem formulation. Try to keep consistent notation.
  • 2 months left: Have a working computational framework. Iterate between paper notation and organization of computer code.
  • 1.5 months left: Have a paper draft. Iterate with Ian.
  • 1 month left: Have a final draft with more refined results. Send to collaborators.
  • 2 weeks left: Continue iterating. Turn in paper if all is going well.
  • Deadline: Submit with confidence.

At 20 hours of research per week, that’s 40 hours per step. My recent sprint shows that it’s quite possible to complete each step in 40 hours, provided the work is done in large chunks. It pays to select a “research day” for making significant progress.

That’s how I should have written the paper, with organization and intentionality and a schedule. Instead I wrote it with adrenaline and panic like a deadline-crazed sophomore. People called me “Don Carter” in high school. It’s time I started bowling like him.


Note: ideas from James Clear are fresh in my mind as I write this. I think the PDF is worth trading your email address for.