log(x) vs ln(x) — The curse of scientific computing

Kamil Tamiola

907

The point about consistent notation is quite true. Though I have to say, this wasn’t as much of a problem 20 years ago, for obvious reasons. I’m a math bod and we used to be quote consistent that log conventionally referred to base-10 and ln was the natural logarithm, with a log subscript b for anything that isn’t to either of those bases (since log subscript e is natural log).

Of course, conversion between logs is also subject to small variations, too. So the approximating polynomial function can often be as much of a problem as the storage of the variable and the discrepancy in the calculation between the two.