Remark -- The answer given below addressed the original version of the posting, which asked if there's an advantage to using \ldots
over ...
(three consecutive dots). At some later stage (May 2013?), the posting was changed to ask a different question, viz., what the best method is for inserting a text ellipsis in a (Xe)TeX document. The OP also provided a new answer at the time.
You asked:
Is there any advantage in using \ldots
instead of ...
?
Yes! If you type ...
you'll get some rather-closely spaced dots. In contrast, with \ldots
the dots are correctly spaced for a typographic ellipsis.
The following MWE illustrates some of the visual differences of what's outpot by ...
and \ldots
. The differences are readily apparent in both text mode and math mode.
\documentclass{article}
\begin{document}
yes and no ... yes and no \ldots\ yes and no ...
$\{x_1,x_2,...,x_n\}$ vs.\ $\{x_1,x_2,\ldots,x_n\}$
\end{document}

Naturally, if your language is not English -- for which \ldots
was designed initially -- you'll want to make sure that the ellipis created by \ldots
conforms to your language's typographic conventions. @egreg's comment suggests that this may well be a concern for the French case. For such situations, you may want to use the commands provided by a specialized package such as ellipsis
or csquotes
rather than the "standard" \ldots
command.
Addendum prompted by a follow-up question by @moose: In LaTeX, there's also the command \dots
. This raises the question, what is the difference between \ldots
and \dots
? The simple answer is: There's no difference in LaTeX. (I'm not sure if this applies to Plain TeX as well; I don't have the TeXbook at hand today...) The definitions of these two commands (from latex.ltx
) are as follows:
\DeclareRobustCommand{\dots}{%
\ifmmode\mathellipsis\else\textellipsis\fi}
\let\ldots\dots
The \textellipsis
instruction, in turn, is defined in the LaTeX kernel as follows:
\DeclareTextCommandDefault{\textellipsis}{%
.\kern\fontdimen3\font
.\kern\fontdimen3\font
.\kern\fontdimen3\font}
Interestingly, there's no kern inserted ahead of the first .
("dot").
...
and\ldots
produce rather similar output (but with the dots apparently raised noticeably above the baseline). Which document class and which specialized packages did you load to create this output? Please consider posting a MWE (minimum working example) that generates the output you show.a…b
anda … b
. In the first case the ellipsis shows that there is a part of a word missing and the latter case shows that one or more words, i.e. a part of the sentence (is), are missing …