Monitoring programmers' stress levels accurately predicts the quality of their code


[Read the post]


Such level of command&control would increase my stress level, even when I not enter a single line of code.





I’m no programer but i’m kinda curious how this tech could be used in game development to monitor things like stress during ‘crunch time’ and whatnot


In a quantified-self, self-managing scenario, where only I see this information, this is tantalising. Like a spell check flagging up areas to revisit, as Cory suggests, or a “go for a walk & see if this problem makes more sense when you come back” alert- a tool to let me know that I’m on my way into one of those regrettable high-stress coding frenzies that I & future maintainers will absolutely regret.

However, the way it WOULD be implemented would be with a project manager appearing at my right-hand shoulder the minute my heartrate escalated. “Is this stressing you out?” they would ask. “Do we need to have more meetings about it? Let’s move to a meeting pod & talk. I’ll invite all the stakeholders”.

One would rapidly learn iron control of physical stress indicators, I guess.


I totes agree. And this got me thinking… a Fitbit like device for techies. And it gives them a mild electric shock when they need to go for a walk.

Anyone want to halp me kickstarter this? :smiley:


just abuse/reuse this thingy : )


Almost missed the obvious one.


Ha! I was a part of this study. The instruments I wore were extremely uncomfortable. I got sore nipples most days. Ahhhh…the sacrifices I make for science.

Let me tell you that I was not concerned about privacy (I was working in an office where I assumed that all of my actions are potentially seen by others).At the time, it was very hard to know if the biometrics device was hooked up properly, if it was reading data correctly, and if I was filling out the right feedback surveys.

I’m glad Thomas and Sebastian got some useful data out of this.


I definitely see this inducing an unpleasant feedback loop, along the lines of “the beatings will continue until your stress levels return to normal”.


dies inside


Stop flagging your code!


Wouldn’t it suffice for a programmer to write a comment to the effect of ‘I am not confident that I thoroughly grok this problem’ above a code segment that caused them to feel that way? Admission of stress is the first step to relief.

Or are programmers generally not self-aware enough to be able to leave such comments?


When things are sketchy (due to imperfect understanding of the problem, or theoretical issues I’m worried about) I usually add a comment to make sure there’s a review or later consideration and to make things easier to debug if we hit a bug later, but comment style varies a lot by individual unless there’s some attempt to enforce a standard (which usually blows). Drawing attention to yourself that way is pretty rough on the ego too, so self-awareness might be why you don’t leave a comment.





Oh, well, as long as it’s for my own good and all … On second thought: no, screw that. I’ll stick to code review, thanks. I’ve known far too many programmers who can’t code their way out of a wet paper control loop and weren’t the least bit stressed out about it. If they don’t care, the monitor simply won’t pick it up, whereas peer code review likely would. (Not only that, but I know I’m not the only one to pick up useful things when looking at the code of other people.)

Some are, some aren’t. Honestly I think there’s a certainly level of comfort you have to reach before you’r comfortable commenting code in that way.

I actually had a recent class where my longtime habit of extensively commenting my code served me well. After the due date of one somewhat difficult assignment, the professor started a class by writing some numbers on the projection. 38 people registered for the class … 15 people turned in nothing for the last segment, 5 people got full credit for it, and the rest … well, some got partial credit, but most were in a ‘pending’ stage as they were using some functions that were well beyond what we had covered in class. He was willing to give them appropriate credit, if they could prove to him that they actually understood the code they turned in for the assignment.

As he mentioned some of the functions people used, I started to be a bit concerned, as I had used … well, all of those, plus a bit more, really, and while I got the assignment to work correctly, I wasn’t 100% clear on exactly how some of these worked together. When I spoke to the professor after class, he chuckled a bit, and stated that I was fine – I was one of the five. Yes, I used those functions, but I had also thoroughly documented every single line where I did.

Over commenting for the win!


you used inline APL in line 325? sleazy.


See also: Taylorism