Sunday, December 18, 2005

Computer science vs programming

Via the CRA, an article that purports to be about the dwindling number of women computer scientists in the field, but is really about much more:
Some computer scientists fear that they may be going in the same direction. They view the dearth of women as symptomatic of a larger failure in their field, which has recently become less attractive to promising young men, as well. Women are ''the canaries in the mine," said Harvard computer science professor Barbara J. Grosz.
The article highlights geometer Diane Souvaine of Tufts, and her work in developing a curriculum that focuses more on the science than the computer in computer science. It makes a point that we in the field are all familiar with, but that I have never seen explained in the media:
Introductory classes zeroed in on programming and other technical aspects of the field, rather than explaining big ideas or talking about how computing can impact society, many professors say. That approach led to a misconception among students that computer science is the same thing as computer programming. Computer scientists say that view shortchanges the field, which is far broader and more intellectually rich. It is applied math and design, they say; it is about modeling human behavior and thinking about the simplest way to accomplish a complex task.
The other point the article makes that I really don't agree with is that a focus on programming and technical aspects of computers is what attracted male programmers (read "nerds") to the field, to the exclusion of females. The implication of course is that if computer science education were focused more on problem solving and "impact on society", that more women would have been inclined to enter the field.

This is debatable. Any higher level "non-programming-centric" approach to teaching computer science would involve heavy dollops of math; linear algebra, graph theory, calculus, probability, geometry, you name it, even if you never ended up doing theoryCS. Math has always had a problem attracting women students, and I don't see why shifting focus away from programming and towards problem solving (which I highly encourage, btw) would make the barrier to entry for women students any lower.

8 comments:


  1. The implication of course is that if computer science education were focused more on problem solving and "impact on society", that more women would have been inclined to enter the field.

    This is debatable. Any higher level "non-programming-centric" approach to teaching computer science would involve heavy dollops of math; linear algebra, graph theory, calculus, probability, geometry, you name it, even if you never ended up doing theoryCS.


    Not necessarily. E.g., take a look at the "Robotics" course mentioned in the article. It introduces a broad set of CS concepts without an overload of math. And at the end of the semester, you have a nice robot that does something concrete.

    Of course, mathematical concepts are (and should be) present. E.g., Minkowski sum makes a cameo.


    Robotics course link



    Posted by Piotr

    ReplyDelete
  2. Actually, the math department at UNC has a really high ratio of female graduate students, from what I've heard. Perhaps even close to 50%.

    I'm female, and I like math. In my high school, it was the girls that were the best at math and physics. You need to readjust your stereotypes some. ;) 

    Posted by Andrea

    ReplyDelete
  3. oh dear. I should have seen this coming. Andrea, your example notwithstanding, it is not stereotyping to observe the fact that math has a hard time attracting women students for higher education.

    The crucial word here is "higher". Jeff Erickson has written about the strange divergence after high school, where from an almost 50-50 ratio of men and women interested in math and science, things start skewing once these students hit college and beyond.

    In a sense, my complaint was phrased in the form of a reduction. If presenting CS in terms of problem solving rather than technical minutiae might help draw students back in, it seemed to me that this merely made CS more mathematical, where there was an existing problem (at the undergrad and higher level) of drawing women students in.

     

    Posted by Suresh

    ReplyDelete
  4. I think making CS more mathematical WOULD draw more women in at the college level. The added math would help flush away the pissing contests among the testosterone-addled programming gods who learned to code in the womb and therefore think CS is easy, even though they can't reason or communicate their way out of a wet paper bag.

    Scare off the assholes, and more women will be interested. And vice versa.

    ReplyDelete
  5. Actually, math does not seem to be as female-deprived as CS is. According to these folks , about 30% of math doctorate degrees went to women in 2003. This is higher than the "around 20%" for CS claimed in the Globe article.

    Posted by Piotr

    ReplyDelete
  6. I recieved my B.S. in Applied Math w/ emphasis in CS in 1997. Frequently, I was the only woman in my math and CS classes. After nine years as a software engineer, I am frequently the only woman on a team, in meetings, or involved in system decisions.

    I do see a difference in the way men and women approach CS and math, and why not. We are different in most aspects of life. Many men approach CS like a video game, usually aggressive and wanting to be number one. Women approach it more like just a job. This comes out in class and in industry, and I think many women choose not to participate. 

    Posted by P.R.

    ReplyDelete
  7. Jonathan Shewchuk1/12/2006 12:10:00 AM

    >> "Math has always had a problem attracting women students, and I don't see why shifting focus away from programming and towards problem solving (which I highly encourage, btw) would make the barrier to entry for women students any lower."

    The page 3 story in the December 2005 issue of SIAM News (not online yet), "Toward a Level Playing Field in Computer Science," discusses how CMU raised the proportion of women in their undergraduate CS program from 8% to 35%. A big part of their strategy is to show that computer science "encompasses much more than just programming," with changes to the curriculum to reflect that. When they surveyed students entering the CS program, they found that "men were more likely to major in computer science as an extension of a life-long passion for computing, ... while the women were more likely to be motivated by applications."

    Another interesting factoid is that CMU discovered that their admissions personnel were "giving a high weight to programming experience," yet "prior programming experience had no correlation with subsequent performance." In effect, they were inadvertently discriminating against women, who are much less likely to have learned programming before college. Changing the admissions process so that programming experience was given no weight increased the number of women admitted.

    The article also mentions that "women account for 30-40% of math majors at top-ranked schools, [whereas] only 15-20 percent of computer science majors at US universities are female."


    Posted by Jonathan Shewchuk

    ReplyDelete

Disqus for The Geomblog