Algorithms+reverse engineering

Everyone interested in software studies and research about algorithms should read this piece by Nick Seaver called "On reverse engineering: Looking for the cultural work of engineers". Based on TheAtlantic's investigation of Netflix's tagging system, the author discusses the consequences of reverse engineering for how we think about the cultural lives of engineers.

Some excerpts that attracted my attention:

"reverse engineering, as both a descriptor and a research strategy, misses the things engineers do that do not fit into conventional ideas about engineering. In the ongoing mixture of culture and technology, reverse engineering sticks too closely to the idealized vision of technical work. Because it assumes engineers care strictly about functionality and efficiency, it is not very good at telling stories about accidents, interpretations, and arbitrary choices. It assumes that cultural objects or practices (like movies or engineering) can be reduced to singular, universally-intelligible logics. It takes corporate spokespeople at their word when they claim that there was a straight line from conception to execution. [...] The risk of reverse engineering is that we come to imagine that the only things worth knowing about companies like Netflix are the technical details hidden behind the curtain. In my own research, I argue that the cultural lives and imaginations of the people behind the curtain are as important, if not more, for understanding how these systems come to exist and function as they do. Moreover, these details are not generally considered corporate secrets, so they are accessible if we look for them. Not everything worth knowing has been actively hidden, and transparency can conceal as much as it reveals."

Why do I blog this? Because it's an interesting argument and practical recommendation for researchers working on such topics. Being interested in the interplay between technical constraints and cultural/imaginary elements, I quite appreciate the point Seaver makes here.