On Being ‘Random Enough’

The concept of randomness is easy to grasp on an intuitive level but challenging to characterize in rigorous mathematical terms. In "Algorithmic Randomness" (May 2019), Rod Downey and Denis R. Hirschfeldt present a comprehensive discussion of this issue, incorporating the distinct perspectives of "statisticians, coders, and gamblers."

Randomness is also a concern to "modelers" who depend on simulation models driven by random number generators or analytic models built using probabilistic assumptions. In such cases, the underlying mathematical model is often an ergodic stochastic process, and the issue is whether the output of the simulator’s random number generator or the observed behavior of the real-world system being modeled is "random enough" to establish confidence in the model’s predictions.

In a sense, this highly pragmatic perspective represents a less restrictive approach to the issue of randomness: if any of the strong criteria described by the authors are satisfied, the output of the simulator’s random number generator or the observed behavior of the system being modeled should be sufficiently random to establish confidence in a model’s predictions. On the other hand, behavior that fails to satisfy these strong criteria may still yield accurate predictions, provided other—less restrictive—assumptions are satisfied.¹ In many cases, these less restrictive assumptions are both simple and intuitively plausible. Their existence explains why many probabilistic models work well in practice even though the rigorous mathematical assumptions these models appear to depend on are unlikely to be satisfied exactly.

Jeff Buzen, Nashua, NH, USA

Back to Top

The Human Side of Computing

I am writing to address the continuing issue of women’s underrepresentation in computer science and, in particular, to Gloria Townsend’s impassioned and well-stated views in "Bringing More Women, Immigrants, to Computer Science" (July 2019) on ways to address this concern. I applaud her efforts and obvious success.

However, I would like to call attention to what might be—in my opinion having taught the subject for over 40 years—one serious cause of this gender disparity: Computer science is far more focused on the technological and informatics aspects of the field rather than the humanistic. In other words, people are spending more time making themselves meaningful to a piece of machinery (sorry, folks, that’s all a computer or smartphone is), deciphering how the software and hardware were designed and implemented.

There is nothing particularly socially positive about this effort. All of us have shared the frustration of trying to speak to a person rather than an automated system whose best feature is to tell you a real person will call you back. (The company would argue the best feature of this system is it saves them money on customer service reps.) Indeed, phishing, spamming, spoofing, robocalling, hacking, and cyberwarfare have added to the negative connotations of the discipline. These undesirable elements steer some socially minded students to choose areas of study that showcase and directly contribute to the upside of the species.

In short, modern computing leaves much to be desired as a human-centered endeavor. Therefore, it is only natural that computer science has become less attractive to people who are disposed (by nature, environment, upbringing, experience, or training) to being socially sensitive and people-oriented. I will leave it open for discussion if this characterization applies (statistically) more to women than men, but certainly the community would do well to emphasize the social good computer science is capable of achieving. We must sell the discipline on its immense social value to humanity.

Steven Minsker, Little Rock, AR, USA

Back to Top

Tight Squeeze

In "Extract, Shoehorn, and Load" by Pat Helland (July 2019), I learned a lot about how metadata is used to ensure the most important pieces of data are translated properly, especially when reducing the size of the data being loaded. I think the analogy to shoes for the way we transform data is accurate and interesting; I never would have thought of it like that. Helland goes into detail about how painful the process of fitting data can be but doesn’t talk about how it could be improved. I think it would be interesting to hear his ideas on how to make the process less difficult.

Mitch Hudson, Cedar Rapids, Iowa, USA

Back to Top

Author responds:

There are a couple of points to remember. First, there’s a difference between the described input and the prescribed output. Second, as I discussed in an earlier article, "If You Have Too Much Data Then Good Enough Is Good Enough" (June 2011), data transformation can be like a meat grinder. When you make a hamburger, it tastes good, but you can’t go backward to the input steak. That’s OK. The hamburger’s tasty.

Pat Helland

Back to Top

Out of Focus

In the article, "The Edge of Computational Photography" (July 2019) Keith Kirkpatrick writes, "… has its subject in focus, and the background out of focus, known as bouquet."

The correct word for this effect is "bokeh." Ask any professional photographer!

Ann Ford Tyson, Tallahassee, FL, USA

Back to Top

Back to Top