It has been five weeks since my last post. In general, they were not an easy five weeks, but The Hacker Within Software Carpentry Bootcamp just wrapped up. Thanks to all of the lovely, sexy, intelligent people involved!
Greg Wilson, of Software Carpentry, gave an amazing plenary session where he described the state of scientific computing. The conclusion was that anecdotally things are very bad, because everyone believes that scientists are anywhere near as effective as they could be with computers. (Let alone the number of computer scientists who can not develop software.)
Greg points out that things are worse than this because we are not even taking quantitative data about how far behind most people are. Worse still, we largely don’t know what the right metrics are, because measuring practical education level is hard.
So with out being too cynical, we don’t know what we don’t know and we don’t know how to find out. (There are some caveats, and things might be improving, but let’s not be too hopeful.) Greg (and now I, as well) hunger for quantitative data on this subject. Philosophy plus experimentally determined numbers equals science, and now we are getting somewhere.
At the symphony tonight, this situation struck me as analogous to almost every issue I could dream up. For instance, this has a fractal effect in science. Say you measure the temperature, you don’t really know the value unless it has some associated error (fractal depth: 1). But then how are you sure that the error is correct (fractal depth: 2+).
Still, the most intriguing corollary is sexual identity (or how one expresses one’s self by who one is attracted to).
I cannot classify who I am attracted to. No matter what I do, there always seems to be an exception to any rule I lay down. (I’ll spare you the juicy details.) However, if you show me someone, or if I interact with someone, I can come up with a decision on whether the subject at hand ‘does it’ for me or not. Moreover, I have no idea how to go about determining a suitable sexy-to-Anthony metric, since all prior attempts have failed. The concept seems fuzzy.
So similarly, to a software developer, give me a person and their code and I can tell you if I think they are good or not. Yet, I find the abstract task hard, unsatisfying, and probably a waste of time.
Let’s take the sexual identity analogy one step further. If we were to sex operating systems, we would probably say that Windows more male, and Mac OS X is more female. Of course, what we are doing here is actually making a woman/man gender assignment.
This makes Linux transgendered!
Unfortunately, the analogy breaks down when you start to talk about solutions in these two spaces. Gender bias probably won’t be solved (in the short term) simply by decreasing the market share of the major players. Additionally, unless you are Richard Stallman, you aren’t just born an open-source-sexual.
I think most people are forced to believe simply because it collapses their information entropy to zero. Sadly, I don’t know how to break them of this habit, or even if I should.