Blog by Sumana Harihareswara, Changeset founder

24 Apr 2022, 22:15 p.m.

How I Thought About an In-Person Conference Choice

Will I go to PyCon in person this year? Today is the day I decided whether to go; I had booked plane tickets and a hotel room, and was running out of time to cancel them without incurring fees. So here's how I thought about it: benefits, risks, and how much I could mitigate those risks.


The biggest benefit to going to PyCon in person would be that I'd be able to participate in the Packaging Summit. I'd represent the work that we're doing through NYU on better securing the Python packaging toolchain, and -- through live discussion -- better understand the state of packaging, what other people and institutions are working on, and so on. There's no remote participation for this event, though I've been assured that folks will take thorough notes to share afterwards.

Second to that: I'd have ad hoc and wide-ranging conversations with friends, acquaintances, and colleagues I haven't talked with for ages, or have only talked with online. This would emotionally refresh and nourish me, and intellectually I'd benefit from hearing their stories and discussing things I don't yet know I want to discuss. In case there's conflict between me and any of my open source colleagues, live conversations could ease things and strengthen our trust foundations to mitigate future conflict. And conversations might lead to business opportunities. Some of this I can get by other means, pinging them and setting up videocalls, but the medium can get exhausting. And I wouldn't get the serendipitous alchemy that can happen in unplanned groups.

Most irreplaceably: I'd have ad hoc conversations with people I don't know yet. I need this because I don't know what I don't know, and people drawn to the same event but from wildly different contexts might open my world to perspectives and opportunities I've never thought of before. And we might become friends! This might happen via the expo floor, in the hallway, and at mixer and Birds of a Feather events. At some previous COVID-era PyCons, PyCon has provided a virtual meetup function that let online attendees mingle in spontaneous group videocalls, but it looks like that is not happening this year.

If I were to participate in the development sprints then I could make progress on projects I care about, recruit more participants, and have a low-pressure environment to demonstrate and learn what it's like to work with specific other people on stuff -- that last data point is really hard to get in other ways, and can be an important stepping stone to other partnerships.

I'd get to visit Salt Lake City again. I have visited with family but never by myself, and would like to explore it a bit.

I would gather some information if I went to talks, and might be inspired by keynotes. Talks are usually recorded for later viewing, but rarely glitches happen and recordings aren't available.


It is still a bad idea to get COVID-19, and I still don't want to get it. I am vaccinated and boosted, so if I catch COVID my chances of having to go to the hospital or dying are lower than if I weren't (though in February, 15% of adult COVID-19 deaths in the US were among people who were vaccinated with booster). My likely course would be a week of feeling awful and maybe some more weeks of feeling gross. But, as doctor and science communicator Bob Wachter breaks it down,

my biggest worry is that of Long Covid (LC), which I now think of in two fairly distinct buckets:

a. Prolonged symptoms, including fatigue and brain fog, that last for many months.....

b. Increasing the risk of bad non-Covid outcomes over time. Recent studies have shown ⬆ risk of heart attacks & strokes, of diabetes, of brain shrinkage, and of blood clots. .... it is this long-term and uncertain LC risk that is my main motivation to try to avoid infection, even if the likelihood of a severe outcome from an acute case isn’t very high, and the likelihood of symptomatic LC in a fully vaxxed person is real but fairly low.

I also don't want to give it to anyone else, especially my spouse, and especially people who are less protected by vaccines or who have less ability to protect themselves (such as customer service personnel).

Specifically with this PyCon: the timing is difficult. If I catch COVID on this trip and need to rest and recover, there are some specific things in the next few weeks I won't be able to do. For example, Leonard and I have special plans we'd have to cancel. And one of my best friends here in New York has a newborn and I'm one of the people who's on call to help out, and I would have to let her down. If my recovery takes several weeks, I'll have to miss WisCon in late May, which would be particularly bad because WisCon and the Otherwise Award are counting on me being there as the benefit auctioneer.

And the BA.2 subvariant is causing an uptick in cases in the US in general, including in Salt Lake City as measured by wastewater data.

Ways to mitigate risks

My assessment is that the riskiest activities would be unmasked or partially-unmasked social settings indoors, and ill-ventilated social settings indoors.

Other people's practices: PyCon has mitigated risks with its health and safety policy. All attendees must wear masks, and loose cloth masks don't count. And all attendees must be recently vaccinated or vaccinated and boosted. Both these steps reduce transmission risk. However, in the shared areas of the conference venue, another convention will be sharing space with us and they have no known vaccination or mask requirement. And I believe neither of the convention hotels has a mask mandate in common areas such as the elevator.

Flight: I live in New York City and flying to Salt Lake City is, for me at this time, the only feasible way to get there and back. Now that the US has no mask mandate for people on flights or in airports (and has never had a COVID vaccination mandate for domestic flights), I would be depending on one-way masking (I wear a mask even though others don't) to keep me safe, from the moment I walk into the airport in New York City till the moment I get into my hotel room in Salt Lake City. I own a P100 respirator and know from experience that I can wear it for several hours at a stretch without difficulty, but going ~8 hours without eating or drinking anything might be difficult. (N95s aren't as effective, but I could switch to them once the plane is inflight and the high-efficiency ventilation comes on. I have been thinking about getting N99s but hadn't gotten around to it yet.)

Eating: To avoid taking masks off in indoor spaces with other people, I would not eat indoors with anyone else. Or I could decide to only do so after everyone in the room had taken an antigen test that came out negative, which would probably only work with takeout or room service in my or the friend's hotel room. This mitigation would reduce the social benefits of going to the convention since so much social availability revolves around shared meals, and lots of restaurants don't have any outdoor dining options, plus it's forecasted to be a bit chilly this weekend in Salt Lake City. I would reduce to a minimum time spent in rooms with other people, especially crowded rooms, especially when others have their masks off: so, no PyLadies Auction, no indoor lunches, no opening reception, and so on.

For eating, I'd depend on outdoor dining, takeout/curbside pickup from restaurants, bringing my own snacks from NYC or buying them in a local grocery store, and eating in my room or outside (e.g. on benches outside the venue or in local parks).

Ventilation: Thanks to the PyCon health and safety reports, I'm reasonably assured of the ventilation in at least one hotel's rooms. The Salt Palace (the convention center) probably has pretty high ceilings. The worst moments for ventilation would probably be in the airport, boarding and leaving the plane, sitting on the runway, and maybe in crowded moments at the venue. I've bought an Aranet 4 carbon dioxide monitor (Aranet 4 discount coupon), which can be configured to update its reading once a minute, on the theory that CO2 levels are a reasonable proxy for unfiltered human exhalation and thus COVID danger levels. A high CO2 reading would prompt me to go from an N95 mask to a P100, or to leave. I'd also want to bring extra AA batteries for the monitor.

To avoid indoor crowds, instead of spending a lot of time in the venue, I'd regularly go into the hallway, find people to have conversations with, and invite them outside for walks or seated chats outside the venue.

Testing: One way to mitigate the risk would be for the conference to mandate at-the-door testing, every day, for each attendee. PyCon is not doing at-the-door testing. So I'd do frequent antigen testing myself. If you have US health insurance, you are entitled to eight free over-the-counter antigen tests, per month, per person in your household, covered by your health insurance. You can submit receipts for reimbursement -- or, big pharmacy chains like CVS and Walgreens have the ability to put it through your insurance like a prescription, so you pay nothing out of pocket. You have time to go get your April tests between now and PyCon; bring them with you to Salt Lake City.

If I caught it: As the New York Times suggests in "A New Wave of Covid-19 Is Coming. Here’s How to Prepare.": make a plan to get antiviral drug treatment if you need it, because taking Paxlovid and molnupiravir early mitigate COVID symptoms and (probably) long-term effects:

"For the pills to be most effective, you need to start taking them within five days of the start of your symptoms, so it’s important to have a plan for getting a prescription and knowing which pharmacy can fill it"

Test-to-Treat is a program happening now in the US: "Through this program, people are able to get tested and – if they are positive and treatments are appropriate for them – receive a prescription from a health care provider, and have their prescription filled all at one location." The Test-to-Treat locator says there's a Salt Lake City location: Fourth Street Clinic, Wasatch Homeless Health Care, 409 West 400 South, a 15-minute walk from the convention area. However, this location is not open on weekends. There's a CVS at 5557 West 4100 South, West Valley City, UT 84120 which is a 20-minute cab ride away, and which looks to be open on weekends -- but then you have to endanger the cabdriver (which you can mitigate by opening the windows and wearing an N95 mask).

And, as the NYT points out, "Make sure you have extra funds or plenty of room on your credit card in case you need to extend your trip to recover from Covid."

Upon getting home I would isolate in some fashion for maybe 3-5 days, probably partially in a hotel and partly in our apartment. I'd probably test every day with an antigen test and go for a free city-provided PCR test a few times during that period. And if I test positive then I'd pursue antiviral treatment and see if Leonard could go stay with friends or in a hotel for a bit. (NYC used to have a program to give free hotel rooms to people with COVID or household members who didn't want to catch it, but that's now over.)

I'm not going

While writing this post, I've cancelled my hotel room and air ticket. The patchwork mitigations I could employ to protect myself and others from the most-transmissible-ever edition of COVID-19 would substantially reduce the social benefits of the con, and be pretty exhausting. I am choosing, with significant regret, to decline one irreplaceable experience so I can save my risk budget for a few others (especially serving as Otherwise auctioneer at WisCon and helping my friend with her newborn).


Sumana Harihareswara
25 Apr 2022, 8:15 a.m.

Another item on the possible mitigations list: improving ventilation in one's hotel room or in shared spaces through user hacks. (And here's wording in case you need help asking conference organizers about ventilation.)

As I mentioned on another social media platform, on the topic of the long run: The trend seems to be that we're figuring out ways to reduce the risk of COVID being fatal or serious, and so I hold out hope that this trend will continue -- the vaccines, monoclonal antibodies, Paxlovid and molnupiravir, etc. And some institutions are getting serious about ventilation (though I'm less optimistic about that as a trend). So I figure that the safety of specific activities will wax/wane in bursts, as the arms race iterates...

And: I use the calculator to help me understand what situations are safer or less safe, but the calculator has not yet been updated for the BA.2 subvariant of Omicron so it's underestimating risk overall.

Sumana Harihareswara
25 Apr 2022, 13:36 p.m.

The Test-to-Treat program also lets you get Paxlovid via making an online "e-clinic" telehealth appointment through CVS ( which is a valuable alternative!