The author of an interactive visualization of dynamics of an infectious disease (inspired by the WaPo article) writes:
You may have heard about "case fatality rate" in the context of COVID-19. This is not the same as mortality rate. "Case fatality rate" is defined as the number of recorded deaths divided by the number of confirmed cases, when both measures are taken at the same point in time. Whereas "mortality rate" is defined as the likelihood of eventually dying from a disease once you contract it. CFR is an important number for understanding a disease as we're learning about it. But the ultimate number we care about is the morality rate...
He references Encyclopaedia Britannica for the definition of case fatality rate, and a Lancet article when he talks about mortality rate.
The article in the Britannica, admittedly, deviates from the common definition of case fatality rate by introducing the time into the definition, but does not support his definition of mortality rate. It says:
Case fatality rate is calculated by dividing the number of deaths from a specified disease over a defined period of time by the number of individuals diagnosed with the disease during that time; the resulting ratio is then multiplied by 100 to yield a percentage. This calculation differs from that used for mortality rate, another measure of death for a given population. Although number of deaths serves as the numerator for both measures, mortality rate is calculated by dividing the number of deaths by the population at risk during a certain time frame.
A CDC book on epidemiology does not support his definition of mortality rate either, and does not include time in its definition of case fatality rate. Rather, when referring to a death rate over a specified time, is uses the term death-to-case ratio, which reappears in other epidemiological textbooks as well.
And yet, incredibly, in the Lancet article he references, they use the term mortality rate in the sense of case fatality rate (number of deaths over the number of incident cases). No peer review for correspondence in the Lancet, I heard.