Causal

Events A, B, C, D, E, and F in a diagram with their causal links. We want to only see that A implies B, but really there are a lot of confounding causes.

“We will start by ignoring reality and pretending everything else is constant except A and B.”

23 Jun 2021

Folders

The folder structure for my data. It's a bunch of folders with very bad names: OldData, PreTweaksData, MaybeUsefulData, FinalData, OldFinalData, TestData, FiguresData.

The fun is when you get back to a project after putting is aside for a while, and figuring out how each piece works together to produce your plots!

21 Jun 2021

Jury-Rig

Left panel (In Theory): "I'll just whip up a small library to take care of my problem." Right panel (In Practice): "The last five hours may have been fruitless, but I'm sure I can jury-rig *this* library for my needs!"

Most of my programming is just a sequence of trying to modify existing code for my needs, before finally acquiescing and writing my own code.

18 Jun 2021

Dopamine

Graph of dopamine versus time for a scientist. Small blips are false alarms, and the breakthrough is a very prominent spike, but the dopamine crashes right after it.

Technically, it’s a Dirac delta distribution.

16 Jun 2021

Compute Time

Two bars in a bar graph. The short one represents the jobs that launch and run without errors on the first try, and then a much larger bar represents the the jobs that end after two seconds from a silly typo.

The fun part is when these errors happen on the supercluster that you’ve been waiting forever on to get the job submitted.

14 Jun 2021

Gap

Left panel: "Can you run this simulation for me?" "Sure, I saw this in class." Right panel: (Later) "How do I even implement Equation 1?!" Caption: Coding shows you exactly where your theoretical gaps lie.

And this is why I’m always wary of papers I read which don’t give any simulation details and just present the results.

11 Jun 2021

Name Vacuum

Left panel: A scientific phenomenon depicted as a circle. Right panel: A bunch of papers pinned onto the circle, trying to establish its name. Caption: Scientists abhor a name vacuum.

Quick, use your academic influence to get your friends to cite the name you’ve chosen, making it appear standard in the community!

09 Jun 2021

Dark Magic

A section of a paper on numerical simulations. Authors will always talk about what they plotted, sometimes talk about how they did things, and rarely give details on why they chose the parameters they did.

The number of times I’ve wanted to know the details of a simulation and was given nothing is truly impressive.

07 Jun 2021

Follow The Evidence

Scientist follow the evidence. Okay, maybe not that evidence.

I mean, clearly that evidence just isn’t good, so I’ll ignore it.

04 Jun 2021

Broken

Regular people are devastated when their work is broken. Scientists are happy to have science break, unless it's their own work.

I need to call Customer Service to the Universe and demand an explanation. Clearly I should have been right!

02 Jun 2021

To the past To the present