Good day Señor K Caj Doog. I came across another World Cup related data set via WASSUP (Weekly Announcements Serving business analytics and mathematical Sciences’ Ubiquitous Programs). This data is about passing graphs between players and implications on style and success of play. They consider certain graph functionals which I believe may have come up in your work as well. I had not seen that Freakonomics blog post, but it is interesting that someone else looked at the same data set. Now that the data set appears in more than one blog post, does it count as a standard data set for trying various applied mathematical methods? What are the criteria for becoming a standard or benchmark data set? How about if I posted a mat file of the data here?
Archive for July, 2010
Did you see this recent post from the Freakonomics blog? It seems they analyze the same dataset as you did. There is an interesting quantization/categorization phenomenon going on, which is perhaps not so different from things we are interested in. Another possible explanation given in the comments is that the observed phenomenon arises due to conversion from inches to centimeters.
This whole business of standardized units of measure is interesting and one of my favorite quotes along these lines is about the notion of information rate measured with bits. As Jerry Wiesner, director of the Massachusetts Institute of Technology’s Research Laboratory of Electronics said in 1953, “Before we had the theory,… we had been dealing with a commodity that we could never see or really define. We were in the situation petroleum engineers would be in if they didn’t have a measuring unit like the gallon. We had intuitive feelings about these matters, but we didn’t have a clear understanding.” I plan to write a separate post about standards, but let me riff a little bit on your previous post and comments thereto.
Ronald Kline, who helped me with the aforelinked paper from the 2004 IEEE Conference on the History of Electronics held at Bletchley Park, has written a couple of interesting papers on cyber-things. The first, Cybernetics, Management Science, and Technology Policy: The Emergence of ‘Information Technology’ as a Keyword, 1948–1985, discusses how various communities have used the keywords information and information technology in various discourses, including the cybernetics community (or whatever it has morphed into). The analytical approach of using keywords is based on the book Keywords: A Vocabulary of Culture and Society, which makes the argument that the keywords used to understand society take on new meanings and that these changes reflect the political bent and values of society. I think this argument from the sociology of language has an interesting connection to the use of linguistic arguments in Sanskritic discourse that, as you know, I am rather intrigued by. I think you were using that style of argument as well. The second paper, Where are the Cyborgs in Cybernetics?, discusses the history of cyborgs and cybernetics, putting forth the argument that the research program of cybernetics developed by Wiener and his followers had very little to do with the cyborgs that emerged in popular culture in the 1980s and beyond. Indeed, I think “cyberspace” is like cyborg studies, having very little to do with the cybernetics of Wiener. (Kline does mention that old-school medical cybernetics was concerned with ‘half-human/half-machines’ which has carried over into neural prosthetics, which e.g. Ram Srinivasan studies.)
Coming back from that aside on “the prefix cyber- [being] everywhere,” I don’t think Anand’s analogy with traditional thermodynamics is directly apt. If one is to make a thermodynamic analogy, I think it has to be with non-equilibrium thermodynamics (which Sanjoy is of course a big fan of). Although I am not a believer in the strongest forms of the so-called strong programme in the sociology of scientific knowledge, it is very clear that there are too many external (social) forces shaping scientific/engineering research to think of it as an isolated system with an entropic characterization. (In all honesty, I think the science of science is still—to use Wiesner phrasing—without its gallon; indeed there may be connections between science of science and service science.) I don’t have an answer for your evolution of fields of study question, but besides internal scientific/technical considerations, I think external aspects of professional identity and prestige surely play into it. For an elucidating example, see footnote 6 of this.
As a closing thought, I also think working with the cdf features [and survival function features] makes better sense than working with the pdf features. Along those lines, let me post a document about birthdays (rather than heights) that I wrote a couple years ago but never made public before now (except to Emre and Emina), though perhaps you and Appu should have been on the distribution list. I wonder if Anand has further insights into birthdays.