In Praise of Local Bookstores

 Op-ed at The Times.

That's enough for today.  I have books to shelve.

Don't let the perfect be the enemy of the good.

 This principle applied to covid-19 vaccines.  RTWT, it's short.

It's funny, but every time I post "eh, throttling back on this blog, not much to say," something interesting pops up.

Still here, don't worry.

 Remember the 2020 wrap-up post that said "Probably lighter posting.  More of documenting special events, such as WFD and Coffeeneuring in 2021.  Probably less of "hey, cool article."  The world is changing, and it's time to change here too."  OK, this is the lighter posting part.  Don't worry though.  Lots going on here, just not a lot of it newsworthy.  I mean, do you really want to slog through trials and tribulations of changing cell phone carriers (which ate the bulk of my last two weeks, btw)?  No, of course not.  So stay tuned, but be patient.  It's the blah time of year, expect blah blogging.

Seven Minutes of Terror to the Surface of Mars

 Compressed into five minutes of narrated video at today's NASA APOD.  You should go watch this now.

ps: Seems to have worked.

The Call of Marconi

 "When the ionosphere was right, They could talk from continent to continent through the sky; but when the ionosphere was wrong, They could not talk to their next-door neighbor.  Not that They ever chose to do so anyway."  -H.P. Maxim, "The Call of Marconi"

Don't struggle with this – either you get it or you don't.

Modeling, Covid, Herd Immunity, and Are We There Yet?

 Here's a grab-bag of articles and links I've found interesting in the last few weeks:

The Hard Lessons of Modeling the Coronavirus Pandemic at Quanta

University of Washington's Institute for Health Metrics and Evaluation: front page | COVID-19

 One bright data scientist doing his best, using machine learning: Interview | Projections | Path to Herd Immunity

I was quietly playing this game too last spring and summer, using genetic algorithms and GitHub data.  It was... interesting, but not especially enlightening.  Watching the weekly medical reporting cycle add a spike-and-dip to the numbers while listening to the media scream "we're all gonna die!" and "pandemic's over!" twice a week in time with medical data dumps was initially amusing, but mostly frustrating.  I mean, isn't basic data smoothing when an obvious cycle is built into the reporting... obvious?

I guess not.  Anyway, those last three links are the best analysis and forecasting I've seen lately.

Half-Way There?

Article at NPR: Why the Pandemic is 10 Times Worse than You Think

TLDR: A research team at Columbia University is using sophisticated models to estimate the total number of people to date who have had covid-19 in the U.S., as opposed to the more commonly-reported number of confirmed infections.  Their estimate is 36%.  

That is half-way to a reasonable number for the herd immunity threshold.  Also, they estimate that the number of daily new infections is around 40% of the peak in late December.  Pandemic collapse?  I sure hope so, but epidemiologists and numerical modelers have been fooled more than once by this virus.  The best we can say is that, for the moment, the numbers are heading the right way.

The above-linked article is short, and filled with caveats that you really need to read.  This is no time to let up from being careful, and we all need to get whatever vaccine is available ASAP.  However, given the numbers and trends described in the article, the overall news it delivers is quite hopeful.  I wonder why the negative, fearful tone used there.

Now I have to say it: Are we there yet?