New York, NY Wednesday, January 1, 2020
Computing Reviews (CR) started in 1995, when I got a call from my former advisor, Uel Jackson McMahan, a neurobiology professor at Stanford and now also at Texas A&M. At the time, journal articles were not available on the web; everything was still on paper. Jack told me many researchers use reviews when starting new projects, or when preparing to teach a new class. However, the good review articles, the objective ones, are usually out of date. Jack’s lab had an idea: put review articles online and update them regularly. This seems matter of fact now, but as Jack used to say: “The best ideas are the ones people don’t recognize at first; then, when you do them, everybody says they are obvious.”
Jack’s suggestion started a journey. First we put Surgical Index online, a CR-like publication for surgeons published by Joseph Ignatious. It was too early. Most surgeons didn’t spend time on the web and most articles were still not accessible online. Then we started re:focus, a service that took controversial topics—censorship, abortion—and used HTML frames to put opposing views next to each
other; readers could use pull-down menus to switch between about ten essays and get an idea of many points of view.
In 1998, we started working with the ACM on CR, with the aim to turn it into a service and make a prototype for similar services across many fields. The central problem we set out to solve remains the same: so much is published in academic literature that it is hard to figure out what is worth reading, especially if it is a little outside of one’s expertise. There are many good ideas out there—some that aren’t recognized, some in a field far from one’s own. If an objective quality filter was combined with strong concept mapping, knowledge discovery could be speeded up and communication between different academic groups enhanced.
CR’s approximately 1000 reviewers are the quality filter. They pick or are invited to cover articles and books in computing and write short reviews that include background, explanations, and critical evaluations. No one reviewer is truly objective, but 1000 different points of view is objectivity.
Now, 20 years later, in a world of atomic swaps and ubiquitous online reviews, post publication reviewing of academic literature is still in its infancy. Perhaps this is just the nature of academics, where anything published can affect a career, or perhaps it is because no one has figured out the key, that thing in the interface or data that makes reviews irresistible and integral to the computing community.
CR is in the beginning phase of automating reviews and word-of-mouth recommendations. In 2020, we aim to improve the selection of reviewed items, increase the number of self-assignments, make reviews more interesting to read, and develop more reviewers, especially from outside the US. These are modest goals, but very important to the integrity and success of the service.
Still, we need to do more. We need to redesign CR to better help readers find the best new ideas in computing. This may require a new type of system and a complete redesign, or it may be more modest, that is, tweaking what we already have. To figure this out we need your help. You, our readers, are the cutting edge of computing. Tell us how you would change or redesign CR to make it more relevant to your work. No idea is too far out or different.
Together, we can make Computing Reviews into something “obvious.”