Thursday, February 26, 2009

Fighting over shiny toys

A recent article in the Guardian extols the wonder of The Algorithm:
And since computers are increasingly dominant in our lives, algorithms are increasingly important - and nowhere is this more apparent than on the internet. In the online world, mathematical analysis isn't just important: the algorithm is king. Everywhere you turn online, companies are using algorithms in their quest for success. From Google's search results and Apple's music recommendations to Amazon telling you that "customers who bought this item also bought ... " algorithms are at work.

The article itself is pretty tame, a kind of knurd version of Bernard's article. What's funny is that the last line in the article, 'Mathematicians rule' got some people into a tizzy.

Two letters appeared in the Guardian following this article:
Bobbie Johnson (Go figure ..., 23 February) highlights the ever-increasing role the mathematics of algorithms plays in our daily lives, including Google's page ranking. "Mathematicians rule!" concludes Johnson. So a reader inspired by your article may seek to contact an expert in algorithms in the mathematics department of their local university. In this, the article will have misled them, as expertise in this area is to be found predominantly in departments of computer science and informatics.

Bobbie Johnson describes algorithms as "jealously guarded mathematical recipes that increasingly dictate how we lead our lives". What he's actually describing is operational research - the discipline of applying appropriate, often advanced, analytical methods to help make better decisions. Executives in every kind of organisation - from two-person start-ups to FTSE 100 leaders - are using OR to structure their problems, unlock the value of their data, model complex problems and make better decisions with less risk and better outcomes.

Of course, the first letter comes from the faculty of the CS department at Edinburgh, and the second from a member of the Operations Research Society. I wait for all the data mining and machine learning enthusiasts to start complaining next.


  1. On a similar note I remember when Netflix had an open request to help improve their recommendations algorithm. There was a contest to help improve their movie recommendation metric.

  2. Actually, I am quite happy leaving it the way it is in the original article. When I meet someone at the proverbial cocktail party and tell them I teach in a computer science department they immediately assume I am an expert in programming languages or databases or some other area that I am not directly involved with. So, it is easier to explain to people that I use mathematical ideas to solve problems - and they get it. I have at least once used the analogy of numb3rs to explain the general idea of what I do.

  3. Here are some definitions of Operations Research I found on the interweb:

    - research designed to determine most efficient way to do something
    - the application of scientific methods and techniques to problems of decision making
    - Application of the scientific approach to complex problems arising in the operation and management of large systems of people, machines, materials ...

    So, by these definitions, algorithms is definitely part of OR. Of course, under these definitions, almost any study of efficiency of anything is considered OR.

  4. I can see it from here.

    Joe who owns a factory goes to ask a complexity theory guy about a problem. The complexity theory guy works for a month and the result comes back: "your problem is in the class XYZ".


  5. This zealot professor from Edinburgh should remember that Knuth is also a mathematician, as many of pioneers in CS are. I think the letters are extremely childish. You do not see real computer scientists and mathematicians saying similar things like "Algorithms are mine!".


Disqus for The Geomblog