,
23 tweets,
5 min read
Read on Twitter
Why is it that so many people have a hard time coding in front of others? How do we end up viewing it as such an intimate activity? Why do we feel so exposed and vulnerable when others try to observe?
This is a something I'm trying to understand better because it's not a problem I have. It's one of the biggest things I see impacting interviews these days. We moved from whiteboard coding to observing people code directly. This is the next major challenge.
This is my current theory as well. But I'd like to hear other thoughts. I don't love this theory because it rests on categorizing coding as a largely creative endeavor that is specific to the individual doing it.
(For the record, people standing over your shoulder while you write emails happens all the time in my experience.)
In my experience, many people hesitate to be directly observed well into their career as an engineer. It's not just about interviews. Ask anyone who tries to push pairing culture into an existing team.
Just to be clear, I absolutely understand the phenomenon of being nervous and taking longer in an interview. That can and should be corrected for by the interviewers. It shouldn't be a dealbreaker. But it doesn't mean you don't have to be observed.
From the perspective of someone who creates interview exercises, one of my biggest annoyances is that employers are creating unreasonably large exercises but still chopping up interview sessions by the hour. If it takes longer, give people more time.
I end up talking to people about exercises that start simple and expand. Ideally your exercise will allow an inexperienced person to show you something, but also it can keep going in order to push someone who is pretty senior. It is *hard* to do this well.
I can feel the tension in this thread. People are bringing all of their traumas into answering my question. That's understandable. But that's also kind of what we need to address. What are the things that make this activity feel traumatic.
We can't lean on past experiences either. New programmers have this same hang up before walking into their first interview. Anxiety about being observed is instilled early and is very persistent. I'd rather talk about it as a solvable problem rather than an inevitability.
Let me offer a different perspective on this. Anyone who's a singer or plays an instrument knows that people will ask you to demonstrate your ability at the drop of a hat. They have no problem asking you to perform on demand. So it doesn't happen with all creative endeavors.
If you're trying to join a band or a chorus or an orchestra, you have to sing or play your instrument as part of the audition. You can't just say "I'm better when there's no one watching".
There are a lot of analogies trying to pin down how to think about the activity of creating software. But they all fall down at some point. I believe that coding is its own activity in many ways. It has aspects of many other things we understand, but in a unique combination.
Many people read my question as pushing people to get comfortable being observed. That's not really what it's about. Instead I've been asking myself the serious question of whether we actually have any objective basis for evaluating engineers. I'm not sure we do.
Some of y'all are clearly making up what you think it's like to be a musician just so you can say my analogy was bad. That's fine. But it's showing me that what y'all think is important about being an engineer is also largely made up.
I believe this is one of the core issues we are circling around. There are many people who have decided there's nothing to learn from observing engineers directly. I believe there is.
I've also been talking about how as discipline we are bad at *teaching* people to be better engineers. I believe that this phenomenon squares directly with the fact that we insist on coding being a solitary exercise. There are few opportunities to invest in active learning.
I should also spend a whole tweet acknowledging something that has come up a lot in this thread. A lot of people do a shitty job at being patient and giving people constructive input about their coding. These people are mostly men. That has done a lot of harm to this process.
A lot of y'all are avoiding letting other people observe you because it feels bad. That's entirely understandable and it's not something I'm gonna argue against.
I wanna add a couple of other observations that might be controversial here. I have implemented take-home exercises. But it requires people to come in and talk through it as part of the interview. These sessions do not go better for those that are nervous.
I have found a strong correlation between people too nervous to code on-demand and people too nervous to talk about their code in-depth in an interviewer setting.
From the perspective of an employer, I believe the take-home exercise mostly results in people spending more time and doing more work. But it doesn't give me that much more signal to make a hiring decision.
The question of how to effectively evaluate software engineers is much harder than we give credit for. I'm not sure continuing to tweak the interview process is the solution.
