This is something I address in the article. I point out that at times e.g. C++ was the only choice for a given problem. Just like Assembly may have been the only solution at times due to limited memory and processing power.

However the point I was trying to make in the article was the bad languages had been picked all through computing history even when better alternatives have existed. Often this has been for political or marketing reasons.

The point I was trying to drive home is that the a language being widely used in a domain does not necessarily point to merit or quality. It could be for legacy reasons, politics, available talent or any number of things.