16

If the unit is the final character in a sentence, the sentence-ending dot is not kerned correctly.

MWE:

\documentclass{article}
\usepackage{siunitx}
\usepackage{mathpazo}

\begin{document}
\SI{}{\electronvolt}. And V.
\end{document}

The eV will have no kerning where the single V does. Can I tell siunitx/pdflatex to use kerning or do I have to apply kerning manually?

Mico
  • 506,678
BandGap
  • 461

2 Answers2

16

The way that units are typeset in siunitx means that they are 'hidden' from any kerning: they are inside a box, which then has various math mode switches. That's required in order to be sure you can control the font weight (bold is tricky to deal with). As a result, you will have to kern such cases by hand. (As an aside, it's arguable that the units are 'mathematical' and so should be in math mode anyway. You'll find that $\mathrm{V}$. also does no kerning, so even without the complexities of siunitx this is still something you may have to worry about.)

Joseph Wright
  • 259,911
  • 34
  • 706
  • 1,036
  • 3
    Nor there would kerning with $V$. either. Also, there's never kerning between characters from different fonts. – egreg Jan 23 '13 at 17:54
1

It looks worse if your units are something like \si{\metre\second^{-1}} kerning after superscript

followed by a comma or full stop. It feels to me like there should be no more than a hair space between the end of the (rather long) superscript and the punctuation, it looks like more than the space between letters to me, so doing it manually would be (a) tedious, but (b) nicer looking. As has been said before, $ms^{-1}$ would have the same effect.

However, you couldn't add too much negative space, in case you have a footnote after the punctuation (or a reference using superscripted reference numbers), or it would look rather like the output of $ms^{-11}$ with a stray comma underneath.

This feels to me like the limits of automation - a macro to automate this would need to know too much.

Chris H
  • 8,705
  • 1
    Units should never be kerned. Not only it is against the conventions it is misleading and annoying. Take your example, is it per miliseconds or meters per second (if kerned) ? How would you know if a A comes after the dot whether it is amperes or the start of the next sentence. Lower case a is even worse. – percusse Aug 07 '13 at 16:50
  • @percusse I'm not referring to kerning the units themselves, but the unit vs the following punctuation. If a dot is used between units to express a multiplication, it's raised (centred I guess), rather than on the baseline like a full stop or decimal point, at least in the publications I'm used to. your point about /ms vs. m/s when using ^{-1} for per is a good one, and I might have to give it some consideration where the unit isn't instantly clear from the context. – Chris H Aug 08 '13 at 07:30