The Danger of Black-Box Abstraction

When I was a graduate student, if you were going to convert some data or something like that, you would write the FORTRAN code to convert the data yourself. That’s how you would do it. Now there are these programs. There are these windows and you click. I .nd with my students all the time, they don’t know why something isn’t working. I’m like, well, did the data convert properly? Open the .le and look at it. It is so black box and it is going from the time when you knew how the data was converted, because you wrote the code to do it yourself, to you don’t even open the .le to see if it is full of zeros or not. So there is a very big disconnect.

Griffn began her graduate training in the late 1980s when trust in one’s computational tools was associated with familiarity with their underlying code. Younger scientists are increasingly comfortable with black- boxed simulations. They grew up with personal computers that did not come with programming languages. They grew up on computer games that offered interactivity without transparency. Unlike a previous generation, they did not program their own games. When these younger scientists work with screen molecules, they are more likely than their elders to give themselves over to feeling in the grip of new materiality.