in https://thereader.mitpress.mit.edu/freeman-dyson-and-me/ Jeremy Bernstein writes:

We began to write letters. I saved his, all written in an almost calligraphic script. The first one is dated 25 November 1958. And the last one I have is dated 26 December 1990. We then stopped communicating by letter and email took over. I have dozens of emails from Freeman. Some of them are very interesting, but they lack the human touch of an actual letter. Can you imagine Lord Chesterfield’s “Tweets to His Son”? Some of the letters are technical. One of them, dated 21 August 1981, begins, “Sorry my friend you goofed.” I had written a book review for the New Yorker that had dealt with maps and had carelessly described the Mercator projection — which projects the spherical Earth onto the plane of a map — as “linear.” Dyson writes, “Mercator’s is not a linear projection but a logarithmic one. So far as I know there is no geometric model for it. Precisely that was the originality of Mercator’s idea.” He then goes on to provide a mathematical proof, and what is typical, he goes on to say, “The really extraordinary thing about Mercator’s projection is that it gives useful maps up to very high latitudes. This is possible only because the latitude scale is logarithmic.” He shows it works well, up to a latitude of 85 degrees, and remarks that the “map includes more than 99% of the globe. No geometric projection could possibly do as well as this,” and then adds, “It is a pity that schools do not teach geography anymore.”

He was of course right, and my error also slipped by the New Yorker’s vaunted checkers. The interested reader can find the details on the web.

– z 2022-10-25 10:59 UTC</div>