Using \colon
gives correct spacing for a map $F\colon C\to D$
in one direction. But writing $L\colon C\rightleftarrows D\colon R$
for a couple of maps in both directions (e.g. for an adjunction) gives disturbingly asymmetric spaces near first and second colons (like if D were a map from R, not the other way round). Probably, it should be fixed (e.g. some other command should be used) — but how?
1 Answer
amsmath
defines \colon
as \nobreak \muskip 2mu\mathpunct{}\nonscript\mkern-\thinmuskip{:}\mskip 6mu plus1mu\relax
. So how about the following?
\newcommand*\cocolon{%
\nobreak
\mskip6mu plus1mu
\mathpunct{}%
\nonscript
\mkern-\thinmuskip
{:}%
\mskip2mu
\relax
}
-
works in simple cases (so I've accepted the answer), but try e.g.
\[\sin\colon X\rightleftarrows Y\cocolon\cos\]
— the space between \cocolon and mathop is too big Commented Aug 25, 2010 at 18:01 -
@Grigory M I found we can easily obtain what you want, by adding
\!
, to be ...{:}\!
...– Lao-tzuCommented Apr 16, 2023 at 8:30 -
\cocolon
! It should be just\lon
(and the opposite of the coefficient ring should be ...).\colon
preferred over:
? I just checked a random sample of books in my office and the overwhelming majority uses the symmetrical spacing of:
. The exceptions (among the 15 or so books I checked) are Atiyah-MacDonald, Besse and Kobayashi-Nomizu. I have to admit that I never sat down to think about this before.\fcn{f}{A}{B}
that expands tof \colon A \to B
. I bet there are people who would say this is easier to parse in the source thanf:A\to B
, though I don't know that I'm one of them.