Try it in the Car

In learning audio engineering, a common piece of advice is to listen to your works-in-progress in the car. Nominally, this is because your car’s stereo probably isn’t going to sound as nice as your fancy studio monitors, and you’ll notice a lot of problems on those that you wouldn’t on the nice speakers.

But there’s something else: when you’re working in your studio on those nice speakers, you’re focused on the project, engaged in the work with all of your attention. I think part of the assumption with listen to it in the car is that you’re also driving, giving it partial attention while your brain is otherwise engaged in another activity. I’ve found there are problems I’ll notice with that frame of attention that are difficult to notice with full my full attention at the mixing desk.

So it is sometimes with software. Many of us who create things people are supposed to interact with via their computers or phones do so on large screens, new phones, fast computers, a reliable internet connection, and with deliberate attention. We create and evaluate our creations in the best possible conditions.

And then we’re often surprised to hear when things don’t work well for people on older equipment, on smaller screens, on unreliable internet connections, in situations where they can’t give their full attention to a task that probably doesn’t deserve their full attention. We’re surprised because we haven’t calibrated our expectations towards reality

• that people still use fifteen-year old laptops, ten-year old phones, and that computing devices have resource limitations

• that internet connections can be flaky, slow, or nonexistant, instead of a fast and reliable

• that people use small screens, or for accessibility reasons have font sizes turned up to the point where the screen may as well be small

• that people may be distracted, stressed, or in a different mental state than pure calm professionalism

Can you expect someone to use your program in the same environment and mental state you test it with? Are you sure? How well does it work on older devices? In a coffee shop with flaky wifi, on the edge of good cell reception? Can people do the things they need to when they’re tired or stressed or inebriated?

Just as how audio engineers use tools to simulate the important qualities of bad equipment, on the web at least there are tools to simulate things like smaller screens or slower networks, but I don’t think they go far enough, and they’re only available to web pages; they also can’t simulate alternate headspaces, like being under time pressure to accomplish a task.

The collective we that make software don’t seem to check these things anymore, and it shows. Try your software in the world sometime: borrow a friend’s old tablet, go to the library and use the computer there, and give yourself a deadline to finish your task and consequences if you fail. You might be surprised how many problems are made obvious by the experience.