Alternatively, you can work with x^0 = ct instead of x^0 = t.
Re:Type Systems (2004) [pdf]
by adrianratnapala on Friday September 01, 2017 @12:01AM
(#15146023)
Yes, and and that a much more common convention than the one I favour. They do it partly because it keeps the factors of c out of the visible equations -- but I like my those cs.
Another reason is that people instinctively feel that the elements of the coordinate tuple all have to have the same units. But that's not necessary when you have a metric around. Indeed keeping irregular units is helpful because it prevents you from accidentally forgetting to multiply through by the metric.
But I preffer to keep that stuff in, even when doing pretty abstract stuff. Essentially because I wanted the type annotations. For example my favorite convention for the Minkowski metric is:
That way your vectors are just $(t, x, y, z)$, where $t$ has units of time and the rest of space. Your covectors do come out oddly as $(c^2t, -x, -y, -z)$, but that's what makes them covectors.