Turtlesoft Blog

A couple weeks ago, Cornell University had over 80 active infections, and it was looking pretty dire. Then, the students got scared, and more careful. Last week there were 29 new cases. This week only 5. Someone was hospitalized for a few days last week, but it’s down to zero again. Still no local deaths.

Cornell has already run about 90,000 PCR tests since classes started. Students are tested twice weekly, and staff once. It’s probably not cheap. But, there’s a lot to be said for testing everyone: it nips outbreaks in the bud, and also acts as a reminder that the situation is serious. Having same-day turnaround also helps.

Meanwhile, I’m still doing research for a Coronavirus risk estimator. The consensus now seems to be that most infections happen through the air, so I’ve read a few dozen research publications about droplets and aerosols. The best is a 1934 paper by W. F. Wells. Here are a couple of its charts:

The first shows what happens to droplets that leave a person from talking, coughing or sneezing. Anything smaller than about 140 microns (.14 mm) evaporates, and turns into a floating aerosol. Anything bigger falls to the ground within a few seconds. What that means for Covid-19 is that there are three basic risks:

1. if you are close enough to someone, you may inhale one of those bigger droplets while it’s still falling. That’s what the 6-foot rule is all about, and the advice to cough into your elbow.
2. After the big droplets land, you can touch that surface, then transfer virus into your eyes, nose or mouth. It’s the reason for washing hands, and not touching your face.
3. If you breathe air, you may inhale those small dried-up droplets, which gradually mix into the entire room volume. This is where HVAC comes in. The risk for any space depends on the number of infected people inside and what they are doing, minus air changes and filtration. It’s also why masks are so effective: they block droplets both coming and going.

The second chart from the Wells paper explains why respiratory diseases are more common in winter. People are indoors more, which is half the problem. Even worse, the air is heated and dry, so more droplets evaporate, float around, and end up in noses and lungs.

When working with droplets and aerosols, it’s easiest to do everything in microns (symbol µm, aka micrometers). A micron is one thousandth of a millimeter. It’s about the size of the biggest tobacco smoke particles, a medium-sized bacteria, or the smallest pollen grains. PM2.5 pollution is 2.5 microns and smaller (the most dangerous size because it gets into your lungs easily). N95 masks filter 95% at ¹/₃ micron size. They hit 99% for both smaller and larger particles. 3M says that’s because the bigger particles are heavy, and ram into a fiber. Smaller ones are extremely light, so the fibers suck them in by electrostatic attraction. ¹/₃ micron is the sour spot in between. A coronavirus is about ¹/₁₀ of a micron.

A few researchers have measured the amount of Coronavirus in mucus and saliva: results range from 12 million to 36 billion virions per cc. It’s simple math to translate that to the amount of virus in droplets of different sizes. As it turns out, a one-micron exhaled particle only has a 2% chance of containing a virus, even at the maximum rate. The bigger droplets that dry up and then float are worse. The maximum size that evaporates in humid air (97 microns) will contain somewhere between 3 and 17,000 virions. The maximum in dry air (172 microns) has 16 to 96,000.

As those droplets lose water, they shrink down to roughly 10 microns diameter (about average pollen size). They become a tiny glob of mucus proteins and passengers, light enough to float for hours, easy to inhale. By the math, those are probably the most dangerous.

Dennis Kolva
Programming Director
TurtleSoft.com

There are two colleges here in Ithaca. Both planned to have classes on campus. Ithaca College (the smaller one) changed their mind a few weeks ago. They will be online-only for the Fall semester. Cornell went ahead with their reopening plans, and classes started on Wednesday. The Vet School laboratories are testing all undergraduates twice a week, which sounds great in theory.

Three weeks ago, there were 7 active cases of Covid-19 in the county. A week ago there were 19. Yesterday, there were 70. Most cases happened because students had parties without masks or social distancing. Who could possibly have suspected that might happen? /s

Maybe everyone will get scared, and change their behavior. More likely, there will be enough cases to trigger an automatic shut-down, per NYS regulations. We’ll see.

Meanwhile in the rest of the US, the State R_t tracker shows wavy curves for every single state. Today the majority of states are positive. Sometimes the majority are negative. It seems to vary on a few-week cycle. Back in April I compared state responses to skidding on icy roads. Because of the feedback delay, it’s easy to lose control and end up in a ditch.

Luckily, that isn’t happening with Covid-19. Most states seem to be converging on a R_t value close to 1. That’s probably the ideal growth rate, as long as the case count is low: the best balance between health and economic activity. I guess it also applies to ice and snow: most drivers in the North eventually figure out how to slow down the feedback cycle and stay on the road.

Globally, infections are also at a steady state, with about a million new infections every 4 days. Much of Europe is starting to see early stages of exponential growth, again. It’s going to be a long haul.

I’m still working on an Excel spreadsheet that calculates Covid-19 risk. There are many studies with useful info, but nothing that translates directly into hourly risk. It will require some assumptions and guesswork to get it calculating accurately.

Covid-19 risk is mostly a matter of HVAC. The amount of virus you inhale is equal to the number of people nearby, times the % that are infected, then divided by the volume of air and the number of air changes per hour.

At least a dozen case studies have been published: cruise ships, a Seattle choir practice, a Maine wedding, church events. I have been using Google maps and other sources to estimate building sizes. Air changes is totally a guess.

The biggest uncertainty was expressed best by Dr. Gregory House: “everybody lies”. Folks don’t want to miss work, or they really need a pack of cigs, so they get into public space even though they are shedding virus into the air. Odds are good that they don’t wear a mask. They avoid testing and don’t get into the official data, so the math is more difficult. Everyone’s lives are more difficult.

Dennis Kolva
Programming Director
TurtleSoft.com

I dropped out of Cornell in 1971, bought some land, and started doing odd jobs. One of the first was at minimum wage, for an old geezer who hired me to put roll roofing on his gambrel barn. He sat in a wheelchair at ground level and talked me through the whole project, step by step. It was all done on old wooden ladders, with a few patched rungs. I was terrified the entire time. Thinking back, there was good reason to feel that way.

Right now I’m re-roofing my own house. I bought some good fall protection for the job, but then decided even that wasn’t enough. So, now there is a comfy scaffolding platform at the eaves, with safety rails. It’s almost like working at ground level.

These days, the risk that scares me most is not falling and dying. Something less could easily be worse: serious injuries, massive hospital debt, chronic pain, permanent loss of mobility. It’s also a matter of comfort. As a builder/remodeler I often worked on roofs with just a safety rope around my waist. Back then I woke up all the time from falling dreams. Now that I do construction more safely, they are very rare. It’s nice to sleep well at night. Less working at odd angles = less back pain. Thank you OSHA.

BTW someone mentioned the unsafe angle of the orange ladder. It’s a support slide for 17 foot roof panels, raised with a rope winch via a pulley at the ridge. Humans use the 3 small ladders inside the frame.

The same risk calculation plays out for Covid-19. What scares me the most is a long hospitalization, then chronic lung damage and piles of debt. Even a mild case still means two weeks of quarantine, boredom and worry. Masks and social distancing are inconvenient, but much less so than what they prevent.

Many colleges that reopened are seeing outbreaks. Cornell and other local schools still plan to open on-site, so I am very worried about the next few months. It inspired me to start building a Covid risk estimator, using Microsoft Excel. It will help for personal decisions.

The process is similar to construction estimating: break things down into smaller pieces, calculate each of those, add them up. For construction, the components are scaffold setup, square feet of metal roofing, skilled labor. For Covid-19, it’s risky environments, and time spent in each. Add up the construction costs and you get a dollar amount. Add up life in 2020 and it’s an estimate of inhaled virus particles, and odds of getting sick.

Turtlesoft got its construction estimating data from our own projects. Covid-19 will be trickier. There’s a fair amount of useful data out there, but it does not translate so easily into virus per hour.

A basic problem for any estimating system is accuracy. Our construction unit costs were calibrated by matching them to actual costs for our remodeling projects. Then we refined them with feedback from users. For Covid-19, I’ll run it on a few case studies like the choir practice in Seattle, and tweak the values so it hits on the nose for those. County case counts and cell phone location data from the Social Distancing Scorecard will also help, especially from back when nobody wore masks.

This is not a rush project, but it ought to come together over the next couple weeks.

Dennis Kolva
Programming Director
TurtleSoft.com

Upstate New York was doing great for a while. From mid-June until July 9, there were only 1 or 2 active cases of Covid-19 in our county. Most surrounding counties were equally low.

I think people started to think the problem was all over. They visited out of state, and had Independence Day parties without masks or social distancing. As a result, Covid-19 surged back. This past week, the number of active cases in the county has ranged from 32 to 39. It’s almost as many as during the peak in late March. Surrounding counties also have spikes. Oops.

At the peak in March, there were 18 people hospitalized with Covid-19. That dropped to zero, most days in May and June. This week saw a maximum of 4, but it quickly went back to zero. There is much more testing now, so March and April probably had hundreds of hidden cases.

Fortunately, still no deaths in Ithaca and environs. Some nearby counties have experienced 10 to 60 fatalities, especially where the disease hit nursing homes.

On a national scale, the State Rt tracker is starting to show interesting wiggly curves for most states. Growth rates go up, until there are enough cases to make people panic and become more careful. Then growth rates decline, until people breathe a sigh of relief and go back to their old ways. Rinse, repeat.

Globally, cases are exponential again. The number of new daily cases is increasing with the same curve as cumulative cases. That’s how exponentials work.

This disease is not fading away just because the weather is hot. It probably will get worse in the Fall. Many vaccines are under development, but getting a well-tested one that’s both safe and effective is still many months off. So the big question is, how to have some semblance of a normal life as long as Covid-19 is still around?

The IHME Covid Projections site makes predictions for most countries and all US states. It now estimates deaths and infections up until November 1, calculated with and without universal mask adoption. The difference is huge, especially in places that are currently hit hard. Overall I think masks are probably the cheapest and easiest way to reduce transmission, and still carry on economic activity that is as close to normal as possible. There are other options too, but none better than blocking those pesky snot droplets right at the source.

Personally, I wear an N95 mask a lot. Definitely indoors in places with other people. Also outdoors when within 6 feet of anyone who is well-masked. I give 20 feet to anyone bare-faced, or wearing a chin-warmer. I saved seven N95s from a box left over from lead paint removal, and labeled one for each day of the week. They all have an exhaust port, but I put tape over it as a courtesy to other people. After 4 months of use the straps are starting to get ragged, but they still work. Hopefully they’ll last until replacements are back in stock.

Many people around here wear masks, and/or give other people plenty of distance. Unfortunately, there are also plenty of Covidiots who don’t give a crap. I asked one non-masked jogger to give me 6 feet as he charged at me on the sidewalk, and he just ran up close and laughed ha ha right in my face. So now I listen for footsteps and scramble out of the way. It’s safer to walk in traffic, rather than close to idiots/narcissists/sociopaths/whatever.

It’s really too bad that the US recommended against face masks for so long. Too bad that leaders didn’t set an example, for so long. The Feds and most states still haven’t figured out effective messaging, to get people to wear them (and wear them properly). I would suggest pushing a conspiracy theory that Bill Gates has cameras everywhere, and masks are the only way to escape surveillance. Its software uses nose and mouth, so make sure both are covered well.

The only good news is that the US approach is going to be great for epidemiologists. There are 50 different states doing different things, and getting different results. Thousands of cities and counties with assorted rules, and varying compliance rates. It’s going to be incredibly useful data.

Meanwhile, I am still stripping old asphalt and adding new metal roofing on my house. The project just hit 25% completion. I’ve never assembled scaffolding before, nor worked with such long sheets. It’s quite a challenge to remove 3.5 tons of decades-old shingle without creating a mess.

So far I have made almost every possible mistake. I guess that is pretty much how I’ve always learned construction. Or programming. Or anything else in life.

Maybe that’s what the USA also needs to do: make every possible mistake. Then the next pandemic will be easier.

Dennis Kolva
Programming Director
TurtleSoft.com

I spent a couple days with Excel, futzing with John Hopkins data for Covid-19. The result is charts showing per-capita infections for the 90 most populous countries (confirmed cases per 100K people). They divide up fairly neatly into 5 groups.

First of all are countries that contained the virus quickly. Except for one city in China, their hospitals were not overwhelmed. Life is close to normal for them, now.

This group is about 27% of the global population, mostly in East Asia. That region had a scary experience with the SARS epidemic, so they knew just what to do for the Covid-19 outbreak: quarantines, contact tracing, testing. Local lockdowns when those weren’t enough. Almost everyone wore masks in public, even before the pandemic. They do help.

Next are countries with a low rate that is still growing. These are about 20% of global population. Almost all are Third World or close. That means the low numbers could be from lack of testing, rather than lack of disease. Australia (thicker line) is the outlier. They contained the outbreak for several months, but Covid-19 has recently started to increase again. It’s mid-winter in the Southern Hemisphere, which may be the culprit. That will also be the theme for other countries, later.

For the next three graphs, the vertical scale is compressed by 5x compared to the previous two.

The “getting serious” middle group has exponential growth, with infection rates that are just starting to be dangerous. Columbia (thick blue line) and Mexico (thick red line) are the worst. India (thick green line) is lowest per capita, but it may be under-reported. These countries make up about 36% of global population.

The fourth group had scary rates of infection in March to April, but they are mostly contained for now. Covid-19 snuck up on them, but all these countries responded well. All are First World countries, with 5% of global population.

Then there are the 15 countries already facing a serious impact from Covid-19, with infection counts that are still increasing. You might say these are the basket cases.

Worst of all is Chile (black line), which currently has 1,527 confirmed infections per 100,000 people. That’s over 1.5% of their population. It made the graph too tall for WordPress to handle, so I had to cut it off at the top. Peru (green) is #2 with .91% infected, Bolivia is #11, and South Africa (purple) is #12 and growing fast. Argentina is also exponential, but just below the cutoff I used. All five of those countries are in their winter seasons. That probably is a very bad omen for the Northern Hemisphere, looking ahead to Fall.

The next most bad-growing-worse countries got that way because of politics. USA (#3 red), Brazil (#4 yellow), Saudi Arabia (#6 orange) and Russia (#7 gray) all have autocratic rulers who started out denying that Covid-19 would be a problem. When proved wrong, they still didn’t do much to help fix things. Sweden (#5 light blue) tried a bold experiment, and didn’t shut down. They ended up with more infections, and an economy no better off than the rest of Europe.

All this stuff is just numbers, but it represents many people’s lives. Looking at that steep red line for the US makes me very angry, and very sad. How did we fuck up so badly? But I’d better stop now.

Dennis Kolva
Programming Director
TurtleSoft.com

New York State has been reopening gradually, with a minimum of 2 weeks between each phase. Ithaca entered Phase 4 on Friday, along with most of Upstate NY. Dine-in was allowed in Phase 3, but many restaurants are still takeout-only, or completely closed. All gyms are still closed, as are most bars. Large public events are still forbidden, and masks are required for all indoor public spaces. Outdoors, mask use seems to range from 0% to 70%, depending on location and time. I think it’s declining.

The R_t tracker site currently shows 33 states with increasing infection rates, and only 17 decreasing. New York had a low number through most of April and May, but it gradually inched up. Today it is exactly 1.0. If it stays there, the number of daily infections will remain the same indefinitely.

Since the current number reflects conditions from a couple weeks ago, it is very possible that NYS has reopened too quickly. The ideal balance between safety and economy probably was back at Stage 2 or 3. We will find out about that in a few weeks.

At least Covid-19 is close to contained around here. The county health department is doing almost 1500 tests a week, with only two of them positive on average. Nobody has been hospitalized here since May 29. Nobody local has died yet. Conditions are worse in the bigger cities within a few hours drive, but still tolerable. Even New York City is inching back from the brink.

Sadly, some other states are starting to become very dire. Most of the US reopened too quickly. Governors are starting to crack down, but they will soon find out what happens with exponential growth and a time lag. For example, right now the estimated R_t for Florida is 1.4. That means cases will almost triple before any impact shows, from remedies started today. Florida already has 9500 new daily cases, so it probably will get worse than New York City was in April, and soon.

It’s like watching a slow, enormous train wreck. Too bad it’s not just a train, but the whole frigging country.

Meanwhile, the big question locally is what happens after Cornell reopens this Fall. They just issued a 97 page report about it, but it leaves many details dangling. I think a lot of their plans ignore what students in their late teens and early 20s are like. The local economy is very dependent on the two colleges, so it’s going to be a very interesting ride.

I am pushing 70, so it makes sense to continue social distancing and mask-wearing, even though risks have declined. Most likely that will continue for at least another year. Masks suck, especially in hot weather, but they are a lot more comfortable than a respirator.

To cut down on exposure, the Turtlesoft office is currently open just one day a week. We probably will keep that schedule for most of the summer. Maybe longer. User support is still available by email, and usually by phone.

I spent the stimulus money on scaffolding. So I’m stripping 3 layers of ancient asphalt roofing off my house, and installing metal. It’s also a chance to finish up the exterior painting a bit more safely, and hang new gutters. After that, we’ll see. If QT seems plausible, at some point Goldenseal Pro will consume much more time again.

Dennis Kolva
Programming Director
TurtleSoft.com