Thanks for the replies,
I still think the quoted article is wrong to suggest that mach meter errors are caused by ISA temperature deviation. I suspect that the errors in early mach meters would be most likely caused by imperfectly compensated position errors (mostly in the static pressure).
A Mach meter, whether mechanical or EFIS, needs inputs of total pressure (pitot) and static pressure. I think temperature deviations will not affect the indicated mach number.