# Should a negative kern be added after a degree symbol if followed by punctuation?

Not sure why I have not noticed this before, but spacing between a degree unit and punctuation seems strange as shown in red. Should a manual \kern be added (as shown in blue)?

## Code:

\documentclass{article}
\usepackage{xcolor}
\usepackage{siunitx}

\begin{document}
{\noindent\bfseries Default:}

If the sum of the interior angles of a polygon is \textcolor{red}{\ang{180},}
then the polygon is a triangle.
The sum of interior angles in a triangle is \textcolor{red}{\ang{180}.}

\medskip
{\noindent\bfseries With \verb|\kern-2pt|:}

If the sum of the interior angles of a polygon is \textcolor{blue}{\ang{180}\kern-2pt,}
then the polygon is a triangle.
The sum of interior angles in a triangle is \textcolor{blue}{\ang{180}\kern-2pt.}
\end{document}

-
You avoid the problem by saying The sum of interior angles in a triangle is $\pi$. ;-) –  egreg May 24 '13 at 8:34
Related Question: Typesetting longitude or latitude followed by a comma. –  Peter Grill Mar 12 at 20:10

• it should be noted that strictly speaking your question pertains to one particular font only (Computer Modern Roman 10, I suppose). Whether or not a pair of glyphs is strangely kerned is, of course, font-dependent. As far as CMR is concerned, I agree with Przemysław: CMR's kerning between a ° and a , seems fine to me; it is the ›corrected‹ version that looks awkward.

• Linguistic or cultural context might also be a factor to consider. Punctuation spacing habits differ from country to country. In France, for example, several glyphs tend to get spaced significantly looser.

• If you found ° that , look too loosely kerned, you'd have to start checking the many other glyphs that are floating way above the baseline as well: superscripts, quotation marks, apostrophes, primes...

• microtypographic issues like this tend to look different in isolated, lab-like form under the magnifying glass on screen vs. in real-life context printed on paper at reading size. It is the latter medium that judgments (e.g. about kerning) should be based on -- as that's what most fonts are designed for.

• for kerning adjustments (not for pair kerning though), microtype's kerning feature can serve as an alternative in some cases.

• how about other fonts? Out of the four fonts I ran a rough test on, it's only Minion that's kerned a bit too tight in some places, particularly around the " (and famously so), but the ° + . I find okay across all fonts. It is, rather, the space between the superscripts and the preceding characters that appears rather small across all fonts. But again, this looks more dramatic on screen than it does in the wild.

-
this is a rather subjective area. when the element preceding the punctuation is part of a math (or math-y) expression, leaving the spacing alone is fine with me. but if a superscript is really a footnote marker, then i much prefer to move it outside the punctuation, to avoid two adjacent blobs of space. regarding punctuation and quotes, as i've said elsewhere, put the punctuation where it is most logical -- the point is that what is being read should make sense. –  barbara beeton May 24 '13 at 14:41
+ 1 regarding the footnote markers -- I almost always put them after the punctuation too (even if the footnote is related to only the last word of the sentence, and I've been criticized for that a few times). But ^^this wasn't supposed to be a model for good punctuation practice. :) –  Nils L May 24 '13 at 14:48
@NilsL: Can you post the code for this? It would be good to know which fonts are which. Also for the punctuation following a quote, isn't that supposed to go inside the quote? –  Peter Grill Jun 6 '13 at 4:52
> punctuation following a quote that's language-dependent. This is the German way (inappropriately) applied to English ;) The fonts are: Latin Modern, Minion Pro, TeX Gyre Termes, Miller. –  Nils L Jun 6 '13 at 8:01

Exercises 18.30 and 18.31 from the TeXbook suggest that we need no kerning:

18.30. $\sin18^\circ={1\over4}(\sqrt5-1)$.
18.31. $k=1.38065\times10^{-16}\rm\,erg\,K^{-1}$.


We have correction neither before an equal sign nor after -1 in superscript.

-