> For example recursion, pointers, closures, and class hierarchies.
> My suspicion is that one or more of these concepts is more of a barrier for women than men.
Seems like quite a stretch, given that those are all concepts created by mathematics, which women have long excelled in - after all, women were the original "computers" before we had machines to automate calculations.
Add to that that the hard sciences like chemistry are full of women, and many of the concepts they involve dwarf the complexity of concepts like classes and pointers.
What changed for women between the time of early programming languages and C++/Java wasn't pointers and classes, but rather the culture of tech itself moved from academia to business, and big business had a heavy male bias.
> given that those are all concepts created by mathematics, which women have long excelled in - after all, women were the original "computers" before we had machines to automate calculations.
Doing rote arithmetic is nothing like doing mathematics. Doing manual calculations is nothing like being a programmer. The activities have zero resemblance. Also, advanced mathematics degrees (i.e. Masters and above) have the same gender parity as computer science. Undergrad mathematics degrees has a more balanced gender parity, but this is explained by a large proportion of those women going on to become teachers.
Also note that the ratios graduating with a BSc and a PhD are very different. There are a lot of challenges around the PhD program that have kept women out. (Challenges like, "long time spent poor during prime child bearing years".)
> Doing rote arithmetic is nothing like doing mathematics. Doing manual calculations is nothing like being a programmer. The activities have zero resemblance.
Understanding the function of pointers, classes, and most programming itself are also nothing like advanced mathematics. Even ML is largely stats and calculus in its technical aspects, not cutting edge math.
There is zero at all evidence to suggest that women are less capable of understanding the concepts behind programming.
>most programming itself are also nothing like advanced mathematics.
This is certainly true. But they do overlap in requiring a heavy amount of abstract reasoning. While "abstract reasoning" is a bit vague, one proxy for this is spatial reasoning which shows strong gender disparities and has implications for science and math ability.
It's extremely vague, and therefore shouldn't be used to justify any conclusions about an entire population.
> one proxy for this is spatial reasoning which shows strong gender disparities and has implications for science and math ability.
Another proxy for it is symbolic reasoning, which is arguably even more closely related to programming (most programming is symbolic manipulation - arising from discrete mathematics). Programming languages themselves were inspired by the study of the structure of natural human languages.
Since women's language skills are often stronger than men's, I might argue that women are better suited to programming than men ... but I wouldn't argue that because that would be using the same fallacious reasoning that is behind the argument you used.
In the real world, most programming and software engineering tasks use a mix of quantitative, spatial, symbolic, analytical, and social reasoning skills. Above all, it takes a structured approach to problem solving, which can be achieved through many forms of reasoning.
Which is why I wish there was better data available about gender differences in ability and interest for programming.
However coming at it from the opposite end, remove all barriers that you can find, and work to make sure your hiring process hires women and men who perform equivalently on the job. As long as those are true, you are probably doing about as well as you can to get the best workforce you can.
Note "who perform equivalently on the job" is a subtle and important criteria. As an example, men tend to have an advantage on standardized tests. Therefore I am firmly in support of such measures as adjusting SAT scores to be better predictors of class performance, and then removing all signs of gender from applications before passing them to the hiring committee. Not with the goal of achieving any particular mix of men/women, but with the goal of getting the best performing class that you can from the available applicants.
There's nothing fallacious about spatial reasoning being stronger in males on average, and math (and presumably programming) being heavy on utilizing spatial reasoning.
It's certainly an open question how much spatial reasoning is relevant, but its not irrelevant and so such data has explanatory power. Of course one doesn't have to be in the top 1% of spatial reasoning to be a capable software developer, but no one is making such a claim. We're looking for factors that help explain women's apparent disinterest in becoming software developers. But its a given that people are drawn to things they're good at and disinclined towards things they struggle with. And so the gender disparities in spatial reasoning is a relevant data point in this discussion.
> There's nothing fallacious about spatial reasoning being stronger in males on average
If you read my comment again, you'll see that's not what I pointed out as fallacious. What I said was fallacious is arguing from just that factor to explain women being lower represented in tech. You can't argue something this complex from any single factor.
> Of course one doesn't have to be in the top 1% of spatial reasoning to be a capable software developer, but no one is making such a claim.
The very specific claim by the OP was that women have more difficulty understanding pointers, classes, closures, etc, and therefore are lower represented in tech.
Also, arguing what is needed for the top 1% is irrelevant because we are talking about gender participation rates across the industry.
> And so the gender disparities in spatial reasoning is a relevant data point in this discussion.
And even more relevant is the significant historical cultural signaling that women don't belong in tech. I'd never claim it's an exclusive factor, but to exclude it from the debate, just like excluding any type of relevant logical reasoning capability, spatial or otherwise, is fallacious.
The very specific claim by the OP was that women have more difficulty understanding pointers, classes, closures, etc, and therefore are lower represented in tech.
You will never see that stated as a claim. It appears as a guess, as a suspicion, and at multiple points I was at pains to point out that I have no evidence for it. It certainly is not a claim.
That said, men are way overrepresented in programming. It doesn't matter whether you look at CS departments of self-taught programmers, they are overwhelmingly (though not all) men. I am arguing that this fact is not in itself proof of discrimination.
Of the claims that I have made, one is that in my anecdotal experience I personally witnessed a lot of people in the early days of the web start in the same role, but the ones who followed a track to "programmer" were overwhelmingly men. When I personally asked women who both had and hadn't taken that path, they told me that discrimination was not a factor in their decisions. Actual responses, "You guys never pushed me away from programming, but I want to make things pretty." "Sure, there have been a few jackasses, but there are jackasses everywhere. Most of you are very welcoming as long as I can do the job."
Again, my personal experiences are anecdote, not data. YMMV and all that. But experiences like these are part of why I don't believe that the skewed ratios are the result of discrimination, no matter how many essays I read claiming otherwise. (Essays whose only real data point seems to be the fact that there is a gender ratio.)
Seems like quite a stretch, given that those are all concepts created by mathematics, which women have long excelled in - after all, women were the original "computers" before we had machines to automate calculations.
Add to that that the hard sciences like chemistry are full of women, and many of the concepts they involve dwarf the complexity of concepts like classes and pointers.
What changed for women between the time of early programming languages and C++/Java wasn't pointers and classes, but rather the culture of tech itself moved from academia to business, and big business had a heavy male bias.