Chesterton’s Invisible Fences
I am well aware of the principle of Chesterton’s Fence, i.e. the principle that ’no reforms should be attempted until the justification of the status quo is understood.’ Over the years, this principle has served me well, in particular when it comes to software development and academic matters. However, I recently discovered that sometimes, the fence is invisible, so you ignore it unknowingly and wreak havoc.
The exact details are not super relevant here, but here is a brief outline of what happened: when organising the reviewing and submission process for SampTA, the conference on Sampling Theory and Applications, I set up things the way I was used to from the machine learning community. We thus had double-blind reviews and double-blind submissions, as well as a programme committee that was used to review machine learning submission. Little did I know that this was not the way things are usually done in that community. Thus, while my co-organisers and I were well-meaning, we just irritated and confused authors unduly. Even when it came to creating the programme itself, we ignored several prominent invisible fences and let reviewers decide on special presentation slots—little did we know that participants were expecting rather more talks so as to be able to connect to a broader community.
How did this happen? Partially, it can be explained by a breakdown in communications; due to the personal circumstances of some colleagues, I had to step in on short notice, and everyone assumed that I had been briefed about how this community functions. I had not, though. Together with the proclivity of academic committees to sit unmoving until forced to move by a strict deadline, I always received the feedback well after the decision had already been made. Unwittingly, I had thus become a blind reformer, confusing and irritating the community.
All is well now, after some explanatory talks, but the lesson is still very fresh: sometimes, the fence is invisible.