Hacker News Reader

Comment by cimmanom | Hacker News Reader

And this is also part of why gender diversity on technical teams matters.

Seriously, I don't know why this is such a hard problem for people.

"We're building tax software. Should we get a tax consultant to look it over and identify problems?"

"We're building a programming IDE. Should we get some programmers on our design team? Should we maybe eat our own dog food and have our programmers use it internally for development?"

"We're building a screen reader. Should our QA team have some blind testers on it?"

"We're building a product that we're marketing to women. Should women be involved in building it?"

Sure you can get around these issues by doing extra market research, but in basically every other area of development we understand that it is an asset to have your designers and builders be intimately familiar with the problems you're trying to solve. The cheapest, most efficient way to make your design team intimately familiar with problems that women face is to hire women.

It's no different than hiring security experts if you're building serverside software - you want some people on your team to be an early voice of reason; to be able to raise their hand early in a code review or a tech demo and say, "Hey, just so you know that's an XSS attack vector," or, "Hey, just so you know I've actually had periods that last longer than 10 days."

i wouldn't blame the technical team for lack of "social diversity" if a stock trading app lacked an important feature.

I would instead blame it on the people specifying the product features for not knowing their market ( and that usually is someone like a product owner).

Note that i clearly agree that having more people inside the company actually representing the target audience is a plus, but that's just a general rule, it doesn't apply specifically to having more women in tech team ( which i think is a nice idea, if only because tech is an interesting field to work in and women shouldn't miss this opportunity)

So who’s to blame for more women dying in car crashes than men because there were no female crash test dummies?

Don’t you think that having women on the development teams might have helped catch the fact that voice recognition software didn’t recognize female voices before the software ever made it to QA, rather than after release?

About car crash dummies: i know nothing about that field, but i suppose crash test dummies are build to represent the "most average human being", gender neutral. The fact that women are a subcategory that require specially manufactured dummies is i think a scientific discovery. Not something the average person would have guessed... I suppose there are other subcategories ( obese peopl, old people, short people) that probably would benefit for custom made crash test..

On female voices not being recognized i'd like you to show me the article that tell this story, because it seems so obvious that i'm having a hard time believing it ( looks like something as obvious as understanding people speaking with different accents).

But that just the thing. When only or mostly men are involved, “gender neutral” ends up being masculine (like the crash test dummies that had height and weight distributions like average men, not like average humans) because that’s our societal default. “Oh, women are a special subcategory.” No, we’re not. We’re half the fucking population.

Let's take the car dummy example. Do you think gender is one of the most relevant distinction if you'd really want to dig interesting clues from your team, that professionals wouldn't have thought of ?

If i were a team leader honestly i wouldn't care about gender first. I'd be more interested in recruiting people from different weights, sizes, ages, maybe even races, and medical backgrounds first. And then, maybe gender would come into play.

For some problems gender is an important factor, for others it's not. The fact that this "gender" predicate divides the world into two half doesn't make it relevant per say. I could divide between blue eyes and brown eyes, or skin tan, or lactose tolerant, etc, and come up with fairmy big chunks.

Now I do agree that differences between men and women are indeed so important that it's an very interesting criteria in many topics. Although for some reason i don't think that's the argument you want to base your reasoning upon.

Like this one:

https://makingnoiseandhearingthings.com/2016/07/12/googles-s...

AI/ML inherits the bias from it's trainers - this is pretty well accepted.

Maybe there would have been more testing on female voices if there were more female engineers, but that is sort of ancillary solution to a problem of poor product design and testing.

It also fails to address the underlying problem that there are simply nowhere near an equal number of available female engineers as male engineers. Shooting for a goal of 50:50 ratio when that is not anywhere close to the ratio of the available candidates makes it mathematically impossible to not make other sacrifices...for what objective?

Your claim is that there are things that female engineers bring to the table that male engineers do not. My claim is that there is nothing a male engineer can do on the basis of their gender that a female cannot, and the inverse is equally true. In fact, I can't imagine a more offensive argument than saying "we need more female engineers so that our app is better at handling women's periods."

Did I say 50-50. I’m saying we should care about having any. Maybe male engineers and product owners could pay attention to female voices and menstrual science. But they didn’t. It’s much less likely that a woman working on these products would miss those issues.

How would female crash test dummies address the problem of women dying in car crashes more often than men? For any given car, women crashing in that car will die more often than men crashing in the same car, because women are much more fragile than men. Dummies will only let you measure the effect.

It isn't really a case of "more fragile", but different body shapes, different center of gravity, and so on. Not to mention that if you are talking averages, the "average" test dummy could easily be built taller, with more muscle mass and a bit less regular fat (even at healthy weights) than the average woman is simply because men would tip the scale a bit upwards. I imagine these things cause differences in the way folks move in accidents.

Airbags, for example: If breasts tend to get in the way, then the airbags aren't going to be as helpful for women or, alternatively, cause more harm to women simply because of breasts (I'm speculating the example, I don't know it to be true).

Please cite evidence that women are inherently more fragile than men.

> i wouldn't blame the technical team for lack of "social diversity" if a stock trading app lacked an important feature.

I would. If you're building a stock trading app, why on earth shouldn't you hire at least one or two people on your design/development team that actually trade stocks and have experience with the current market?

It's not even a question of social diversity; it's just that you'll catch so many problems if you do - I can't count the number of times at my last job we ran into technical decisions about performance and algorithms where we didn't know what the right decision was because our dev team wasn't familiar with how the software would be used.

You can't user-test architecture, at least not without very long, very costly development cycles and refactors. It's valuable to be able to catch problems early. Even at a purely technical level, having a dev team that's unfamiliar with what you're building is a recipe for disaster.

Why can’t she report it as a usability bug? Instead she complains loudly to her readership.

I guess Fitbit’s next step should be to recruit female testers to avoid Twitter mobs.

Maybe she did report it?

A report may or may not result in a fix. A post like that might accelerate progress there (given some reach).

If the company is big enough, complaining loudly to your readership is often the fastest way to report a bug and to get someone who can actually do something to look at it.

Take a look at the most recent web audio stuff that happened in Chrome. Bugs had been sitting open in the tracker for months before the code got pushed into stable - dev team didn't do anything about it until prominent developers started calling them out publicly on Twitter and news sites started reporting on it.

As a smaller example, I used to do API level maintenance on a software stack. I didn't get to prioritize whatever bugs I wanted - upper management did. I actually advised people who had breaking bugs not to just file something in the tracker, because I couldn't guarantee that I would be able to take time to fix it. On top of filing an issue, I told teams that had serious problems that they should go to a manager and explain that it was going to make them miss a release deadline.

Then I'd be allowed to take time off of the more flashy features that looked good in press releases to fix their problem.

There is at least a decent chance that somewhere on the Fitbit engineering team, someone is reading that article and saying, "finally we get to address this #?&x$ issue because now it's public enough to care about."

> I guess Fitbit’s next step should be to recruit female testers

Yeah, they should, because female testers are more likely to catch problems that will effect females. They should also hire blind testers to catch accessibility problems. They should also hire runners and joggers to test if their tracking is accurate in normal usage.

I'm not sure if you meant this as a joke, having people who are similar to your end users test your product is basically business 101.

This is clearly due to lack of user testing.

This is clearly a lack of research before the sub-par user testing.

Reply on news.ycombinator.com