Did you read the survivorship bias portion of the post? Since it is making points in support of the direction taken here and you didn't respond to them I assume

Did you read the survivorship bias portion of the post? Since it is making points in support of the direction taken here and you didn't respond to them I assume not, but in particular it gives an interesting exasmple of how Elm evaluates feedback from language users (it collects data on the error messages they get automatically using elm/error-message-catalog).