Post #4717 J: Shutdown Diary; Nest; Baseball's Back!

COVID-19 Shutdown Diary

One of the things I meant to discuss in earlier posts is the fact that the clinic my personal physician works through sent out an email or snail mail announcing they now provided COVID-19 antibody testing, which should be covered through most insurance

I was less than impressed by this. As the reader may know, there are basically two types of tests: RNA and antibody. RNA tests are the kind most people think of by testing: do I have the disease? Antibody tests are basically looking for evidence you've acquired some form of natural defense against the novel coronavirus, most likely through past (unknown?) infections (vaccines are another way, but not yet available). Now in most posts what I've found most interesting is though officially only about 1% of the population has been diagnosed (through generic/RNA tests), multiples of that (maybe up to 5% or more) show antibodies, e.g., in one notorious Silicon Valley study.

Now really a key construct is herd immunity. Basically the idea is that a sufficient proportion of people are naturally protected to limit spread of a contagious disease in a population. For example, I had all the childhood diseases when I was young, but measles had  essentially been eradicated by the time my nephews and nieces were growing up. But anti-vaxxing resistence tripped the herd immunity threshold, leading to a resurgence in the US (e.g., spread by people visiting abroad who had caught the disease). There's a freeloader problem; we do need some slack to accommodate people who physically cannot tolerate vaccines for health reasons, but those who refuse vaccination for reasons of convenience or ideology risk public health. Some British researchers argue the COVID-19 herd immunity threshold may be as low as 43%.

That being said,  I can only speculate that employers might be interested in antibody tests want some reassurance from COVID-19 outbreaks in the  workplace, thinking those with antibodies likely won't spread the virus and/or are immune. The problem is, I don't think we'll get to herd immunity in the workplace until presumably safe, effective COVID-19 vaccines are available and distributed, at best early next year by current estimates.

Now, the day after I just published a post on face masks and public policy, I got an email from Delta Airlines (as a past customer), discussing their pro-face mask, once again, referencing the infamous Missouri hair salon case study (re CDC) where stylists caught COVID-19 but none of their masked customers did. (Oddly enough, they also say the stylists were masked (but somehow caught the disease anyway).)  Anyone in a first-year empirical research course will discuss limitations of case studies; in particular, it's difficult to draw general inferences; we often have small nonrandomized sample sizes, the data are often incomplete/imprecise/subjectively-categorized/qualitative (vs. quantitative), treatments/visits may vary, etc. I am not familiar with types of hair salon services, duration of services, the nature and extent of stylist physical interactions with customers, particularly near or in front of the customer's face and breathing passages.What is the nature and extent of information do we have on customers? For example, what if some had been previously exposed to the novel coronavirus? What kind of data are we collecting? Questionably reliable subjective reports on health? COVID-19 tests?  What type and what time period for reporting? (For instance, it may take up to 2 weeks for the disease to manifest and develop; maybe the customer was exposed but recovered by the time of an RNA test, which might be revealed by the antibody test, but if the customer has antibodies, when did they initially develop? Before or during the salon experience?)

It would not impress me if someone more familiar with the case study specifics can answer some of those questions, which I do not claim to be exhaustive. A case study can provide a context driving theory and hypothesis testing under more rigor, standardization of measures, randomness and statistical power, etc. The problem I have with what Delta Airlines is doing is trying to compare the experience in ONE scenario of salon to the context of airline travel. For example, when I've gone to 2 doctor's appointments in recent visits or to the barber shop once, I was only rarely or briefly within a few feet of other people, even while wearing a mask--mostly just a provider or assistant. When I'm on a plane, I probably sit near a handful or more of people within a 6-ft radius and probably longer than I had exposure to a provider on those visits. Note that I am not trying to fear-monger here; I think the probability of getting infected on an airplane flight is probably minimal. And I don't have an issue with people being aware that using masks can be safer than not using them. But keep in mind that most people use masks that maybe deflect maybe 2% of airborne microbes and maybe larger virus-laden droplets from others (sneezing, coughing, talking loudly, etc.)

I think the real story behind the case study really has to do with the duration of exposure and local concentration of local microbes. The stylist who caught COVID-19 from her infected colleague was probably near her for most of the workday; the mask she was wearing did not keep her from being infected. In my recent essay, I speculated  there may be a threshold of microbes (which may vary by individuals) beyond which point the body succumbs to infection, so limiting exposure to any one (non-family) individual may also be part of a proactive strategy.  (Don't read too much into the "non-family" distinction; of course, family members can catch the virus from others and pass it to relatives. But you may know them better and recognize the onset of symptoms.)

Google Nest and Usability Issues

I've discussed some issues I've had with Google Nest in my practical computing blog. I'm not bashing Google here; I'm a huge fan of its products and services (including but not limited to Internet searches, email, this blog, Youtube videos, my Google Hub, Chromecast, Chromebook notebook, browser usage, cloud storage, and my smartphone operating system). The Nest is a smart thermostat that you can program for energy-saving purposes, remote control, etc.

Whereas I have often been an early adopter (e.g., I bought a VCR before anyone else I knew), in other ways I was a late, reluctant one: for example, I have an extensive vinyl record collection, and I moved to CD's when new releases became almost impossible to find on vinyl. In this case, my home had a good manual thermostat, easy to operate; management decided to upgrade their properties with smart thermostats. I was somewhat dismayed in fixing something not broken, without my knowledge or consent. No instructions given on its operation; now it's not that difficult to figure out to use it by playing around with the interface, but I remember one autumn morning when one had to switch from the air conditioner to the heater, having woken up cold: how the hell does this thing work? It wasn't like you had blue and red dials or levers, something we in ergonomics would call a "natural" interface, where I could link to it from other hot/cold mechanism contexts, like operating a shower.

But, and other users may know what I mean, what really annoyed me was the "stuck white house icon" experience; in essence, the thermostat was hosed, not working, and resetting it wasn't as easy as online help or Youtube videos lead you to believe, although I expect I was dealing with a defective one. It literally took me hours the first time to get to a reset screen, and I'm not sure what,if anything, I did to get there. (It's trivial to reset if the interface is usable.) I've had multiple stuck interfaces since then, although the interim ones took only a few minutes to resolve. But then this week, I woke up from a nap to find the temperature over 80 and a stuck white house icon which I was never able to get past over the coming day.

That was bad enough, but I only had what I can describe as a drifting thermostat problem. Basically it wasn't registering the correct temperature. So a basic sample problem was: suppose I had a heat threshold of 65. But the Nest thought it was 68. In reality it was 60. So the heater never cut on. Now I've got French-Canadian blood; I've been able to go out in short sleeves when most people wear sweaters or jackets. But I knew it wasn't 68 in the home. I had to go to Walmart and buy a simple portable digital thermometer. So I went to management pleading to replace my Nest. For some reason, they wanted me to reseed the device before replacing it (which for some reason I hadn't tried). When I did, suddenly the device reset to 60, and my heat cut on. Confused, I subsequently sent away the maintenance guy with the replacement unit, thinking the problem had been fixed. Big mistake. Within a half hour, the Nest thermometer had drifted up to 68 again, when in reality it was 64 from the heating, which had cut off when it hit my threshold. Management ignored my complaints.

At some point I ended up with the Google Nest support people, not to complain about Google Support because I've frequently found most tech support sucks; I've had notorious issues with Oracle as a professional DBA over the years (I remember one duty support manager who was so pissed I had escalated an issue from his incompetent junior analyst, he called my Oracle Consulting manager, promising to help recruit my replacement if my boss would just fire me). You're frequently dealing with people who have scripts to go by and can't cope with problems conceptually. Problems vary, but in one case I vividly remember this guy basically had pulled up pages of knowledge base responses from an ill-specified search, and 90% weren't even conceptually related to my problem. But he insisted on going through each item in serial order, and when I objected to this, he would term me "difficult to deal with" and threaten to close the TAR (problem report). I recall getting it escalated to a more senior analyst, and he resolved the issue in less than 30 seconds. The problem is that I was working for an Oracle customer with real business issues; tech support has their own internal procedures and incentives which don't necessarily dovetail with customer needs.

Another example, a classic Oracle horror story, was when I worked as a subcontractor at MasterCard some years back. I was given direct orders from my supervisor to raise an issue with Oracle Support and get an answer; Tech Support then was initiated by phone call, and Oracle Support was deploying a variation of Support triage: only what they considered high priority issues would get near-term support. They decided my issue didn't merit same-day support; I tried arguing with them and got nowhere. That didn't stop them from calling MasterCard and bitching that one of their DBA's was violating the terms of  "Gold Support" and threatening to cancel their support agreement. (That would never happen, of course; they were making a lot of money off expensive Oracle maintenance fees.) But MasterCard managers don't like trouble-making contractors and jumped all over me; a contractor is one phone call away from being escorted off their property.

That wasn't the end of it. MasterCard had procrastinated on getting me access to voice mail and a network account, so I had to work away from my cubicle. So Oracle eventually calls back on the TAR in question and when I didn't respond to their unknown voicemail, they call up the client to bitch about me again, canceling the TAR  because I'm not "professional enough" to return their phone calls. This kind of bullshit is fairly typical of the nonsense I've had to put up with over the years with tech support in general.

But this isn't Nest support. My first problem was detailing with all their support documentation requirements: what type of Nest device/model are you dealing with, the serial number, maybe purchase details, etc. Understandable from their standpoint, but it's not like I bought the unit and had paperwork. I have decent eyesight, but it probably took me a half hour to make out the tiny characters in the serial number. Seriously, dude? If you need customers to report the SN, why the hell do you make it impossible to read? And the model ID wasn't obvious, not like a nameplate on a car; I had to pull up a screen of Nest models on the Internet to identify the one I had.

I recall there was something they asked for where I had to mount the device via USB and read some text file stored on the device. I don't recall what happened; maybe they needed purchase specifics I couldn't provide and/or couldn't get from management.

So I started up a new case this week with the newly stuck "white house" icon, and the support agent wanted me to mount the Nest, so apparently I could somehow to a reset manually. But this time, for God knows what reason, it's not mounting on either of 2 PC's. I'm checking ports, cables, etc. I briefly see an undrillable USB drive. Google finally seemed ready to process a replacement.

But that means what? I have to wait days in a hot apartment for a replacement? I go back to management. Finally I've got a functional replacement, at least for the time being

Some Entertainment Notes

• WWE. By far, the most interesting storyline is the Sasha/Bayley partnership as Sasha looks to become "Two Belts" (Tag team and RAW) like her Smackdown partner. I don't like the idea of taking the belt off Asuka. Meanwhile Shayna Baszler, who seemed to disappear for weeks, recently reemerged, although the storyline is fuzzy. There were rumors of a possible Rousey return, but still everyone expects for Sasha and Bayley to feud for Bayley's belt at SummerSlam. Ziggler was a throwaway contender for McIntyre; a lot expect Orton to be McIntyre's next challenger, while the Strowman/Wyatt feud is expected to resolve at SummerSlam. I like they finally put the Smackdown tag titles on Nakamura and Cesaro. The Raw feud between the Street Profits and the Viking Raiders has become stale. It's not at all clear what they're doing with MITB winner Otis, who has all but disappeared from TV except for stupid unlikely romantic vignettes with Mandy Rose.
• Christmas in July. HMM is back to normal programming, and it looks there's another week for Hallmark Channel, although it looks like they do a comedy interlude for a few hours after 2 AM EDT.
• Baseball is Back! It's actually been a while since I've watched a regular game (I've usually watched playoffs, the Series and the All-Star game), but I'm starved for fresh entertainment. Very bizarre because Fox seems to be simulating a real crowd. I know some fans watch from Chicago rooftops, but that's not it.