Childcare During the Apocalypse

Parenting is hard in general, let alone during a pandemic. With schools closing, working parents have been forced to adapt to try and juggle their home responsibilities with their jobs. 30 percent of healthcare workers are parents of children under the age of 14 who might have to choose between caring for their children or working.

One innovative solution to this problem is Covid Childcare Co-op Calendar, a schedule creator that working families can use to try and handle childcare responsibilities as a collective. Participating families mark down what days they’re unavailable for and the site will generate as fair of a schedule as it can, deciding which caretakers will be responsible for which days.

The site was built by Malcolm Harris, author of Kids These Days and Shit Is Fucked Up And Bullshit, and a team of engineers including: Darrell DeCosta, Jane Huang, Jackson Flint-Gonzales, and Arseny Semin. We sent over some questions for the team so we could learn more about the goals, setbacks and technical considerations for the project.

**Malcolm Harris: **Most of my organizing work is via the Philly Childcare Collective, so that’s my first thought in a crisis. We can’t do our normal work supporting political organizations by doing childcare for their meetings, so I thought this was a way to prompt people to think about holes in our childcare response.

MH: Surprisingly few, considering how fast we were setting out to act. We wanted to make sure that what we were suggesting (small childcare pods based on pre-existing relationships) wasn’t contrary to best practices, but when I researched the literature on childcare and epidemics, it turns out to be a lack of childcare for healthcare workers that’s the real threat to public health. We talk about “flattening the curve” but we also need to increase/maintain our healthcare capacity, and childcare is part of that.

**MH: **We’re still trying to figure that out exactly! We’ve had users from around the world, but we’re not tracking their usage so it’s hard to know exactly how it’s being used. That said we’ve also had some great responses on social media. For me, it’s partly about technology as rhetoric, how do we prompt people to start thinking about these questions? But we also want them to use it!

**MH: **Definitely. The state has really abdicated its childcare responsibility, starting Zoom classes and otherwise throwing up their hands. For a lot of caregivers, that means a whole additional job they have to be doing all of the sudden. Even for those that have extra time now, that’s heavy, but for “essential workers” it’s even tougher. Meanwhile, millions of children are training for the telecommuting jobs of the future.

MH: We’ve had a ton of people step up to translate the tool, right now it’s available in English, French, Spanish, Portuguese, German, Mandarin Chinese, Hindi, and Greek. From the beginning our users have been international, just like the crisis, that’s the level at which we need to respond.

Darrell DeCosta: I've used Glitch a bunch and have always loved it, both as a platform and as a company. Being able to see changes being made in real time as various people joined in to help made going from zero to something we could use super quick.

Arseny Semin: I had not used Glitch before, but when it was mentioned I went and tried it out and found it really convenient for immediately showing feedback. It started as just a place to share html and [JavaScript] mockups as we got a feel for how we wanted the site to look, but then became the official development site where new changes were tested and presented. It was so straightforward that I wouldn’t even say a “decision” was made to use it. The question really was “why wouldn’t we use this?” when it already worked so well.

Jackson Flint-Gonzales: A fair amount of our UI was actually from scratch. We likely could have saved some time by, for instance, dropping in a calendar from a UI library, but we opted to make our own. We didn’t want or need the hypergloss of Silicon Valley-crafted digital management tools, so politicizing aesthetics meant a little bit of extra work. At the same time, saying we consciously crafted a political posture in our design would be giving us too much credit. Much of the decision to work from scratch came about as a consequence of the absence of tools that serve the needs ours is fulfilling.

There’s this idea that open-source is a huge pile of Legos; you fish through and eventually find all of the pieces you need to make exactly what you’re building. In reality, open-source things have certain end-products in mind when they are made, and if your end-product isn’t something that a lot of people need, it doesn’t get community support. In our case, the end-product that many people need that most resembles the CCCC is employee scheduling software. But the aim of the CCCC differs from that of employee scheduling.

The latter is like an airline; it wants to fill every single open spot given a pool of people, and once that decision is made the pool of people only have a small amount of control over the result. The CCCC doesn’t seek to impose that kind of strictness on its users, and it won’t, because ultimately it’s a convenience tool, not despotic tech that tells you when to go to work. When you have these differences, it trickles down to every aspect of design and functionality.

AS: Yeah haha. I think I literally googled “multiple day calendar selection js” and clicked on a few results until I found the most barebones example that fit the bill. Malcolm had a sort of pitch for the site being “90s” style. Static, an html page and a script file. So that’s what I was working off of. Then you know, bootstrap, fontawesome, stuff to maybe take the edges off some of the buttons. The real “big” component that wasn’t really just available via a simple stack overflow search was the actual algorithm for computing fair days. But fortunately a team member knew a CS professor, who knew of a solver library that he and another team member helped implement to fit our inputs and outputs.

Jane Huang: Some heavy math was involved in ensuring that a fair schedule was generated. There’s actually a whole branch of computer science that’s useful for scheduling problems in general, and we took advantage of that to ensure the schedules generated would be fairly distributed. We also were very keen on never storing any user data, so there’s no backend to speak of. It’s completely client-side; that is, all the code needed to run it is in the user’s browser. The lack of a back-end also made banging out a functional prototype in a weekend a lot easier.

JFG: Our group’s designer chose a color palette that maximizes visibility across different kinds of color-blindness. We are making continuous updates to the app to improve screen-reader and keyboard accessibility.

AS: There was a lot of translation work done by volunteers after the launch, which was very exciting to see. And we had to pull in a [JavaScript] library for changing dates to whatever the locale’s format was. Or else you’d end up with dd/mm mm/dd confusion.

DD: We actually spent a lot of time discussing this. There was a really strong desire to gather info that would help us get an idea of how people were using the App, but we also very much wanted to avoid any kind of tracking or collection of personal data. In the end we decided to forgo any kind of third party analytics and opted to make use of the real time logging provided by the CDN we're using (Fastly) to collect ONLY the following: The referer, rough geo info (city, country, and continent), and what language the user's browser requested so we could prioritize translations. We also log when a user generates or shares a calendar, but we do not capture ANY info entered into the site. All logs have a 7 day max retention.

AS: Yeah this was part of that whole “90s aesthetic” thing. The site doesn’t have a [database], the code itself doesn’t store anything on the user. If you want to store “state”, it can encode that into a url that you can copy paste. You can share the same calendar with others that way, but there’s no id or anything associated with it, we couldn’t track that if we wanted to, and we don’t want to! You won’t “login” to view your calendar, and it won’t remember you when you leave. The only people who see your name are the other people in your coop, because you sent them a url with your name in it.

JFG: We used Discord to chat in real time and, of course, Glitch to share the work-in-progress. It’s hard to say what might’ve been done differently if there were no urgency to get the tool out. We could speculate an alternate history where a childcare co-op said, “Hey wouldn’t it be cool if we had this tool” but that would be a scenario in which COVID-19 didn’t happen in 2019/2020 and make families across the globe take on a huge increase in childcare. That’s really what produced the sudden need for the CCCC.

AS: Honestly, yeah, Glitch mostly! I could write a change and then link the discord saying “hey here’s that thing we talked about” and it was immediate. As for doing anything different, I don’t know, it went pretty well!

*If you like this story, check out this one about a mutating newsletter or this one about a year long creative coding project.