I would have expected \meter\per\second to generate an output m/s. But when using the package siunitx the output is similar to ms^-1. I referred the documentation for the package and there too the latter is presented. What may be the reason for this? Do SI standards recommend it that way? Is the simple m/s obsolete?

  • Perhaps the following serves as a reference: Derived units expressed in terms of base units
    – Werner
    Commented Sep 25, 2012 at 5:27
  • 4
    Nothing has 'changed' in siunitx: this has always been the default. Perhaps you mean 'a difference in the standard settings between siunitx and SIunits is ...'.
    – Joseph Wright
    Commented Sep 25, 2012 at 5:58

1 Answer 1


It is still possible tu use this notation in siunitx with the option per-mode = symbol and maybe also combining it with the option bracket-unit-denominator = false which will allow you to tell siunitx not to use brackets to delimit the denominator.

This notation has be abandonned for a very simple reason which is that it is most of the time ambiguous. For example, an ampere per square meter and per second would be A/m^2s or A/m^2/s or A/(m^2s) ...

For a scientific usage, it is way better to use one notation which is used by everyone and is totally unambiguous, including for a computer. Therefore, the use of powers only was chosen.

Also note that this enables you to use only powers which were used anyway for powers but also for fractional powers (square roots ....).

  • 2
    Pretty much spot-on. Note that the point of siunitx is to be flexible: I have to pick a default. Powers are most general, so are the standard setting. The package will add brackets to prevent ambiguity, but lots of brackets don't make for easy reading :-)
    – Joseph Wright
    Commented Sep 25, 2012 at 5:56
  • Link to a related topic, on parsing inline division on Math.SE. While it is not yet a universally accepted convention, I would argue that A / m^2 s is both unlikely to cause confusion in practise as it has only one sensible interpretation from a standpoint of how people tend to lex/parse mathematical expressions — that being A m^{-2} s^{-1} . Commented Sep 25, 2012 at 12:30
  • @SamuelAlbert +5 for 'including for a computer'. As an engineer I appreciate that we are considering that as a factor. And I digress, but I also find it neat that the first amendment is being extended to machines. Commented Sep 25, 2012 at 15:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .