Are anti-patterns stalling our DEI work?
I recently stumbled upon the concept of so-called “anti-patterns” in computer science. The idea is simple: for large software projects, it is important to build into place structural “patterns” in how the code is constructed (meta-algorithms, if you will) so as to facilitate collaboration, simplify debugging, and ensure the code is readable and extensible. An anti-pattern, by contrast, is a pattern used over and over again to resolve problems despite the fact that experience shows it to be ineffective.
Anti-patterns are so pernicious because they easily become embedded in a project or institution despite everyone’s best intentions. That’s because anti-patterns don’t necessarily stem from malice or incompetence – indeed, they often are the natural result of locally optimal decisions that seem to make sense at the time, but in reality keep perpetuating the same mistakes in the big picture.
Take, for instance, the anti-pattern I have seen operate in many experimental physics groups: the “software-rewrite” pattern. Basically, early on in the life of an experiment, the software for running the experiment is written in a hurry and often patched together from multiple sources. (This is to be expected: at the earliest stages of the experiment, it is perfectly natural to prioritize getting something running.) Eventually, the graduate student or postdoc who wrote the code leaves the experiment, and the next generation of lab members discover that it is faster to rewrite the code from scratch than to try to deal with an incomprehensible (and buggy) codebase. Not wanting to waste valuable experimental time, a lab member rewrites the code in a hurry and takes the responsibility for maintaining it. Eventually, that group member leaves too and the code ends up getting rewritten again…
Obviously, the solution to the “software-rewrite” anti-pattern is simply to have someone take the time to write readable code. Over the life of the experiment, the “software-rewrite” anti-pattern wastes months of time. Yet the anti-pattern happens anyway because the locally optimal strategy at any point in time is to minimize time spent writing code punting the problem to future generations.
Unsurprisingly, anti-patterns abound in the administration of physics departments and in the microcosm of individual lab groups. This is in many cases to be expected – we train physics Ph.D. students to be future researchers, not future lab group managers, grantwriters, administrators, teachers, or every other role we expect faculty to perform! Moreover, the decentralized nature of many physics departments means that decisions are made without a global sense of accountability or long-term effects. As a result, even the best-intended attempts at solving a problem are likely to end up repeating a common anti-pattern. I outline 3 such anti-patterns below in the particular context of diversity, equity, and inclusion (DE&I) initiatives (though the lessons can be extended further):
The “silo” anti-pattern
The silo anti-pattern occurs when a physics department has many valuable initiatives going on, but each is the pet project of one or a few individuals and does not fit into a broader whole. The silo anti-pattern emerges in an environment where individuals are empowered to begin projects, but there is no broader leadership creating a vision for the departments’ efforts and ensuring long-term stability for the initiatives that are started. Essentially, all initiatives get stuck in the pilot phase, falling apart the moment the person who started them backs away unless said person can find a successor. Moreover, the silo anti-pattern inevitably leads to burnout, since individuals who take on a cause bear perpetual responsibility for whatever they take on.
The “whack-a-mole” anti-pattern
As I discuss in this article, the “whack-a-mole” anti-pattern occurs when departments respond vocally and forcibly to individual incidents of discrimination or misconduct without addressing the underlying culture problem. The reason “whack-a-mole” is an anti-pattern is that, by itself, condemning “bad apples” won’t solve a structural problem. If you are faced with have a bucket of rotting apples, it is necessary but insufficient to remove the wormiest ones. You have to get rid of the nasty filth the remaining apples are floating in.
The “robo-committee” anti-pattern
We’ve all heard the expression “the committee is the only lifeform with 9 stomachs and no brain.” Actually, having served on plenty of both highly functional and highly dysfunctional committees, I tend to disagree with this folk wisdom. Diversity and inclusion committees can be highly effective provided they are given a clear vision and the power, resources, and independence necessary to carry it out. The anti-pattern occurs when the immediate response to a problem is to create a new committee without considering how the committee fits into the broader structure of the department as a whole. Since a department’s capacity for service work is finite, the inevitable result is a proliferation of largely impotent committees that merely consume resources.
I have listed above 3 of the most pernicious anti-patterns I have seen at work in physics leadership at the level of the department or research group. Of course, this list is by no means exhaustive -- I’m seriously thinking about writing a similar list for teaching practices that constitute anti-patterns in the classroom! But what I hope we get out of this article is the concept of an anti-pattern itself – a lens through which we can examine institutional processes that seem to be stuck repeating themselves for non-apparent reasons. If diversity initiatives are stagnating despite intense effort, for instance, the problem might very well be that they were (unintentionally) structured as an anti-pattern.
The key to challenging systems ridden with anti-patterns is to avoid blame. Remember, anti-patterns can emerge even from well-intentioned, logical leadership decisions that merely aim to solve today’s problems. We can recognize that all actors engaged in an anti-pattern are doing their best and likely don’t desire the outcome of the anti-pattern any more than we do. At the very least, even if we can’t interrupt an anti-pattern in progress, we can help make sure that lessons are learned when we go back to evaluate what went wrong – else anti-patterns will continue to repeat themselves while baffled physicists watch well-intended initiatives continue to bear little fruit.
