Future of Coding • Episode 75 A Case for Feminism in Programming Language Design In the academic field of programming language research, there are a few prestigious conferences that you must present at to advance in your career. These conferences are rather selective about which presentations they’ll accept. If your research work involves proving formal properties about a programming language, you’ll have their ear. But if your work looks at, say, the human factors of language design, you might as well not bother applying — and thus, not bother pursuing that work in the first place. Why is the formalistic, systems-focused work elevated, and the human-focused work diminished? And what are the downstream consequences, the self-reinforcing feedback loops that come from this narrow focus? In this episode we discuss a paper by Felienne Hermans and Ari Schlesinger titled, A case for Feminism in Programming Language Design. It applies the lens of intersectional feminism to reveal a startling lack of “Yes, and…” in academic computer science, where valuable avenues of inquiry are closed off, careers are stifled, and people are unintentionally driven away from contributing to the field, simply because their passions and expertise don’t conform to a set of invisible expectations. Through heartbreaking personal anecdotes and extensive supporting references, the paper makes the case that there’s a lot of high-value greenfield work to be done, and people who would love to do it — but we will need to collectively identify, understand, and then fix a few broken incentives before it’ll happen.

I paused the podcast to watch the talk at Onward! Unfortunately, since it's embedded in a super long video, it's not immediately obvious how long the talk is when you start watching it. For anyone who's wondering, it's about 50 minutes long.

Fuck, that's interesting. As a person who is cis-male, and who struggles with questions like "why does no one else give a shit about making this power accessible to real people?" I don't think I would ever have connected those dots. Coming at it from a legal domain-specific perspective, I'm starting to realize that I'm frustrated by a culture in comp sci that cares about formal semantics but not user design, and by a culture in law that is more interested in winning an argument than about helping people! I'm with @Lu Wilson. The more I think about it, the more problematic everything I'm doing seems. Like that bears a lot of thought. But it also gives me so many useful ideas. I think I will be able to communicate the value of certain things better, now, by explicitly confronting biases I didn't realize were making people care less. Great episode!

Does anyone know of a language manual or a reference manual for Hedy? I couldn't find one so I started slogging through the tutorial and my brain broke when I saw:

name is Hedy

age is 15

print name is age years old

Is this some kind of macro expansion language like m4? I couldn't figure it out and I couldn't find a reference.

There's quite a bit of writing about the design of Hedy in the GitHub wiki. But no, I haven't seen a language reference either.

This paper does talk about bit about quotes, though not a complete deep dive. Might help make it understandable https://hedy.org/research/Hedy_A_Gradual_Language_for_Programming_Education_2020.pdf

@Ivan Reese @Jimmy Miller Thank you both for the references!

We can. But if we do, then we are left with symptoms and no diseases. If there are differences between genders, and those differences act in predictable ways to determine what is important or unimportant in the field because of imbalanced gender representation, and you want to change what is viewed as important, that is prescriptive: get more gender diversity among the decision makers, and what is considered important will change. The argument is, I believe, that we might have gotten to all those "feminine" innovations even faster than we did.

@Personal Dynamic Media I think it's wrong to dismiss these very serious matters as "weird gender stuff". I'd encourage you to revise.

The argument is, I believe, that we might have gotten to all those "feminine" innovations even faster than we did.

I don't think that theory fits the facts. Most of the research I mentioned occurred as soon as the technology existed to support it. Some of it (Logo, Smalltalk) occurred BEFORE the technology existed to deploy it because they were anticipating the future. To clarify, I'm not arguing against inclusion. I'm arguing against the gendering of research areas.

@Personal Dynamic Media

I don't see how gendering these types of research is helpful.

The root issue is about who possesses decision-making power, what values those people hold, what incentive structures this establishes, and who is boosted or sidelined in their research career as a result. The use of a feminist lens is one way (of many) to identify these power dynamics. Another issue is about gender diversity in the field, and how evidently easy it is to produce a list of important work done by men, and how much harder it is to produce such a list done by women. There is a subtle element of identifying the cultural dynamics and predilections that appear in work done by men vs women, and while I believe that's a real and important factor, nobody is saying that we need to let this be the framework for making future decisions. It's just input we can use to help understand what's going on. If this doesn't make sense, that's okay. We can, perhaps, help explain the thinking, and that might help it make sense. But I will strongly caution everyone that this is a highly sensitive subject. I'm going to moderate this discussion. For instance: I do not want to see ideas being dismissed because they're not understood.

I can understand the confusion if the claim is understood to be that spreadsheets are feminine and type systems are masculine. I don't think that's the claim. I think the claim is that for whatever reason, people at different points on the gender spectrum have stereotypical patterns of behaviour, which reveal different values, and those values are reflected in the sometimes implicit social norms around what counts as good or important in a community when one gender is dominating it. And that other kinds of good things are being paid too little attention to, because the community focus is arbitrarily on a subset of values. It doesn't require the things being valued to be gendered. It just requires gender norms to have different emphasis on certain values. And it's not a simple effect, or mutually exclusive, because things have multiple facets on which they can be judged. Maybe masculinity values competition, femininity values cooperation, so both value team sports, but for different reasons. It's the reasons or the motivations that are gendered. At least that's how I'm interpreting it. I could be wrong, because I'm new to the idea.

I wonder if this is just an example of toxic feminism? I strongly agree with Hermans' observation of "Hard = value, Easy != value", but, I'm not so sure that this can be blamed on gender bias. [I blame this on something else, something like "certified smartness" (IIRC David Ungar)]. I (male) experienced this first-hand when someone (male) told their manager (male) that I must be lying, because I couldn't satisfy the desire to show how to implement a semaphore using my node-and-arrow notation. What I was demoing (live, for real) was just too simple to be useful in practice in their opinion. This is not a binary issue. There's gender bias AND there's smartness bias. It's not one or the other, it's both, conflated together.

@guitarvydas This phenomenon you describe is exactly what the paper refers to, and we explicitly talk about it on the podcast episode. I'd encourage you to look further into it if you haven't already. I don't think it's right to name this "toxic feminism" and I encourage you to revise that

I agree there's weird gender stuff at play here. It's really weird that this slack community is such an overwhelmingly male space and it makes me feel really icky about participating in it as a man. The lack of understanding, self-awareness or empathy that leads to @Personal Dynamic Media rejecting the issue as 'weird gender stuff', @guitarvydas questioning it as 'toxic feminism' and @Karsten Wagner saying they're fearful of even talking to women and black people underlines this and gives extra weight to Herman's argument. That said I also naively find the idea that qualitative = feminine and quantitative = masculine a bit awkward. This point was nicely brought up in the q+a, and signposted in the podcast, along with nice points about how a more fluid conception of gender can help offer solutions which I'm fully on board with. The paper doesn't work as a conclusive work by an expert on intersectional feminist theory offering answers, but does work very well as the start of a massively overdue conversation [1], by an expert on programming language design, taking time off their actual work to make clear how ill-equipped the research community generally currently is to talk about or represent the 'future of coding'. [1] Of course this is not the first start of this conversation, but still feels like a start, targetting people who have been carefully ignoring it.

There is nothing toxic or weird about understanding the relationships between societies and genders. Dismissive name-calling is toxic, though. Where Lu has asked people to revise, I'd ask you to apologize, because your words are offensive to me, and I anticipate to others. Failing to acknowledge and address that offense is not conducive to the kind of conversations I think we want to be able to have, here. If we should be lucky enough to be joined by more feminine and feminist perspectives, we need the space to be one they can feel welcome. Or, as has been said, we risk being another example of the problem Hermans' paper describes.

I largely agree, but I don't think there's been any name-calling here (if you feel I'm wrong, DM me, I want to keep this discussion on topic). To their credit, Paul and PDM both seem to be coming from a place of "I don't understand", not "I don't want to understand", which I appreciate. Yes, "toxic feminism" is missing the forest for the trees, but this is something we should be able to help unpack and explain. I read it as genuine uncertainty about what the paper / talk / episode are arguing for and against. This is good — and I really appreciate how Alex expanded on this rather than trying to shut it down.

I think I understand what you are aiming at but I think it's misguided. I don't think people have been called names. But I think ideas have been dismissed in needlessly offensive ways. I'm the father of a non-binary child. Saying things like "weird gender stuff" in a conversation with non-binary people is not okay. It's not an attack, but it's obviously insensitive. People can be genuinely uncertain and offensive at the same time. I don't think there has been any incivility. But I think we can and should aspire to more than civility. Civil conversations can still be very excluding. Comp Sci conferences are nothing if not civil, and the paper was about how excluding they can be. I'm also not saying what people can or can't say. I'm saying how words have affected me, and maybe others, and asking the speakers to take those effects into account, advising what would be helpful. They get to say things like "weird gender stuff", and I get to say things like "that's offensive, please apologize". If you want to hold space for people, that's how you do it. By letting people express when the space feels hostile, and allowing people to correct for that. If you can't even make the request, then you are protecting the offense and not the offended. And, I have to say, the idea that emotional reactions are "off-topic" has got to be like the definition of toxically masculine. I want all genders to feel comfortable in our conversations, and I don't think that sort of norm is going to help at all. For what it's worth, I think everything so far has been constructive, and healthy. If no one but me is offended, and/or no one cares, that's 100% okay. But I'm too old to give a shit, and too frustrated watching the world around me become increasingly hostile to trans people to say nothing when people are being insensitive about gender issues. If I step over any lines, let me know. But I will tell you if I think the lines are in the wrong place.

Hmm, are there parsing ambiguities here? "weird (gender stuff)" vs. "(weird gender) stuff". Or, "the toxicity of pushing the label of 'feminism' too far" vs. "feminism is toxic". Or "we need to insert more females, because only females are capable of thinking <in a-particular way>" vs. "we need to insert more people who are capable of thinking <in a-particular way>". I think that there are 2 orthogonal, distinct issues here. Branding them both with the same label causes an important issue to be ignored and/or to be misunderstood. We certainly benefit from Hermans' talk and discussion of the issues and observations about a years-long "experiment", but, that doesn't necessarily mean that Hermans' theory is correct. Hermans' did call for more science. The scientific way to check out a theory is to criticize it while keeping in mind that the criticisms might fail to disprove the theory. Here's how I understand Science: 1. perform experiments 2. observe and report 3. posit a falsifiable theory 4. attack the theory in hopes of finding a single data point that disproves the theory, as stated. Step 1 can, also, consist of things like research, performing analysis, or revamping a previous theory. Step 4 is usually done by others who are supposed to have a less-biased view of step 3. Sometimes, step 4 is done by the originator, too. Step 4 is often called "replication" but, IMO, that naming under-emphasizes the fail-fastness of the scientific method. Science is a fail-fast methodology. That's how science results in rapid progress. 1 & 2 are valuable (but, they constitute only a part of science), step 3 cannot be deemed to be "true", ever. If all known attempts at step 4 fail, then the theory remains only a theory. Science is all four of the above steps. If you don't do all four, then you ain't doing science. Stopping at step 3 ain't good enough, IMO. An example: the Michelson-Morley experiment was a step 4. It disproved the then-current theory about ether. It did not actually disprove the existence of some kind of ether. Another example: Ptolemaic Cosmology remained a theory for some 1,400 years. Further advances in experimentation resulted in a different theory about cosmology. The fact that Hermans did valuable work with 1 & 2 does not automatically mean that Hermans' version of 3 is true. In fact, IMO Jonathan Edwards has a different version of 3 and is actively testing it (OnWard! conferences, iterations stemming from Subtext and the like).

@Jason Morris — Completely agree with your reply to my messages. That extra context helps me understand where you were coming from and what you were saying, and I feel quite similarly. With respect to my mention of "on topic", I should have been clearer: I didn't see any instances of name-calling, and I don't want this thread to devolve into "when I said _ I meant ___" posts (like this present one) because that's not going to help us reflect on Herman's work. So if there's any way we can avoid that (perhaps by moving such discussions out of this thread), that'd be my preference. @guitarvydas I don't think the way one parses "weird gender stuff" changes what's hurtful about it. My read (not the only one, ofc) is that it suggests that it's inappropriate to be conscious of gender when looking at certain issues. I'm of the belief that it's unwise to separate the human element out of any of the work we do. The work is done by people, in society, for people, always. It doesn't matter if the work is about people or not; it all bears fingerprints. In all of life and society, but especially in computing, there's a plethora of longstanding, intense imbalances between the genders (and races, etc). To address these imbalances, we need to be mindful of them at every turn, in every context we can. We should look at everything we have made and ask, "why is this the way it is?" in every way we can. Some people (IME, all of them men) might find it ridiculous to ask questions like, "what role does gender play in the merit-based selection of papers to accept to a conference?" But time and again, when people (IME, most of them women) do the work to collect the data, they find heretofore invisible gender-related consequences and influences behind pretty much everything we do. (For instance, I love the snow plowing example.) Now, I need to acknowledge one of my biases that I'm working on: all the above is a very binary way of looking at the matter. I talked about "imbalances between the genders", which strongly implies that I hold a binary view of gender. And, hey, I do by default think of gender as binary. I'm wrong to do that, but I have a whole life of habit to unpack. I'm not far enough along to speak as confidently about it as I can about, say, the imbalances between men and women. In some ways it's as simple as "just don't do that", but to echo Jason, that's treating the symptom and not the disease. So apologies for being a bit clunky about this.

We certainly benefit from Hermans' talk and discussion of the issues and observations about a years-long "experiment", but, that doesn't necessarily mean that Hermans' theory is correct.

Hermans' did call for more science.

The scientific way to check out a theory is to criticize it while keeping in mind that the criticisms might fail to disprove the theory.

Can you point to where Hermans suggested we take a scientific approach to resolving the issues raised in her paper? I don't doubt it's something she could have said, but I want to be sure we're not putting words in her mouth.

Thanks for your detailed answer. I find it quite illuminating. 035749. Hermans' exact words are "... we don't have any theoretical frameworks for programming language understanding or adoption ... if we could have just a little bit more human studies ...". She doesn't use the word "Science" but I automatically interpret what Hermans says to mean that. Is there another interpretation that I'm missing? I seem to agree with everything Hermans says, except when Hermans uses the words "feminism" and "women". Hmm. Yes, the snow plow example is very interesting. How should one couch the rejection I received when I tried to broach the subject of programming simplicity? In particular, I think that node-and-arrow is amazingly powerful, but,,, must not be judged by failureful implementations that began life with synchronous programming ideas. I, also, knew one (now deceased) implementor who swore off of all academic conferences because he was tired of continuous rejection. I doubt that the term "feminism" would apply to him. I don't know how to reconcile this knowledge with what was presented. Comments? Disclaimer: I, also, think that PL is an outdated concept. That train of thought must undermine how I receive information. [I would venture to guess that I've been tinkering with PLs longer than Hermans has been alive].

I think discussing what 'the scientific method' might be a digression so I'm just going to leave this here https://bigthink.com/articles/there-is-no-scientific-method/

I watched the talk (loved it, agreed with everything), and quickly looked at the paper, but did not yet have an opportunity to read it. Some assorted thoughts based on the talk, the podcast episode, and this discussion. Mostly rambling and expressing how I see things :) We live in a society. Our society is patriarchal (and imperial, and colonial, etc.). This leads to a multitude of issues. The situation with women in engineering and PL is a reflection of the situation in general, and has mostly the same root causes. The society defines and prescribes genders, and everyone is harmed by this. The further you are from the ideal “patriarch” that is at the top of the food chain, the more you're harmed, and even cis hetero men are harmed by all of this. Examples where men reject other men's ideas because of simplicity and similar concept — a consequence of this, of how toxic masculinity expresses itself, and is the same idea that this paper (and the one about icebergs) talks about: society look at “hard” stuff as being good and worthy of pursuing, and at “soft” stuff as being bad, meh, and “someone else/anyone else could do it”. Those that are at the top want to keep all that they find interesting and fun to themselves, and anything that is below them — that's for those who are below. Women are expected to care for the family, etc., etc. Add to this all the other issues, with all the harassment people get, where the more “different” you are — the more you get it. Intersectionality, all of that. It is quite logical that people who receive harassment will try to evade it. There is this thing with the leaky STEM pipeline, where even when you try to get more underrepresented folks into the tech, the chances are they'll get bounced back after everything they will experience once inside. This talk made me question my tendency to choose and enjoy hard stuff. When a bias can be defined and exposed, it allows you to take it into account. And while I can't say I ever dismissed the “easy” things, I will try to give more attention to things I do not consider “hard” in the future.

Wonderful podcast, as always. I came here to 👏 @Lu Wilson’s final point about ditching "the spectrum", and share an analogy that helped me understand this better: politics. (Another sensitive topic! But bear with me; it helps because it invokes less essentialism.) People often categorize themselves and others on a left/right spectrum. But if you dig into the details of what it means to be left/right wing, you end up finding a bundle or cluster of viewpoints that are hard to meaningfully tie together based on a strong, coherent ideology. Rather, this bundling of beliefs is due in no small part to sociological forces. And these bundles vary over time and place. (Talking politics with someone from a foreign culture can be a surprising, rewarding, and fascinating experience because of this.) Left/right are a convenient shorthand, insofar as it gives you a concise way to refer to a bundle of beliefs (in a given culture, at a given time!). And yet for any given person, it will fall short, to a greater or lesser extent. For some people, it will fail spectacularly. Something similar applies, mutatis mutandis, for gender and the attributes that are associated with gender. (An undergrad professor of mine ages ago pointed out that when romanticism was dominant, women were described in writings as coldly rational; later, when rationalism was en vogue, women were described as hopelessly emotional.)

You imply this is positive, but I see it less about everyone happily discovering independent thinking, and more about far-right forces and state-sponsored information warfare attacking the culture, solidarity, social contracts etc that hold our societies together. I dunno.. Whenever someone says to me they're 'neither left or right' without fail they then swiftly move through a series of alt-right talking points that they've been drinking up from unhinged social media.

I'm pretty conflicted. It is fascinating; it is definitely not purely positive. I see deep and genuine value in social contracts, commonality of culture, etc. And I also see the serious harm that the prevailing system and norms have had on lots of people (including but not limited to women, racial/ethnic minorities, religious minorities, trans people, and folks who are just plain different). Blowing up structures to free those imprisoned by them leaves you without structures. I've been thinking about this a lot. I wish we had some way to have it all. On balance, yes, I err on the side of freeing people, but I am under absolutely no illusions that it comes without a cost, and that that cost is real, and very much worth worrying about.

For those who read Dutch, or are willing to tackle it with some AI aid, there's an opinion piece by Felienne Hermans on roughly the same topic, but for a much wider audience ("Volkskrant" is one of the main national newspapers in the Netherlands). I have never before seen this topic been discussed outside of academic or tech circles.