# [texhax] Strange behaviour of size of parenthesis using lmodern

Eduardo M KALINOWSKI ekalin at laposte.net
Mon Jul 30 13:24:55 CEST 2007

Morten Høgholm escreveu:
> On Mon, 30 Jul 2007 06:25:23 +0200, Chris Rowley
> <C.A.Rowley at open.ac.uk> wrote:
>
>> =?iso-8859-15?Q?Morten_H=F8gholm?= wrote --
>>>
>>> You should report this as a bug to the Latin Modern authors.
>>
>> Or maybe as an inelegance in the design: better or worse than a bug?
>
> In CMR the numbers are not strictly inside the bounding box and the
> same is the case in LM - except for 4 and 7 where the bounding box has
> been extended so that the character is contained in the top (but not
> the bottom for 7).
>
> Running this example with and then without lmodern shows it quite
> clearly:
>
> \documentclass[11pt]{article}
> \usepackage{lmodern}
> \begin{document}
> \setlength\fboxsep{0pt}\setlength\fboxrule{1sp}
> \fbox{$\scriptstyle 2$}\fbox{$\scriptstyle 4$}
> \fbox{$\scriptstyle 6$}\fbox{$\scriptstyle 7$}
> \end{document}
>
> Quick tests show that for many of the fonts in LM, figures have
> different heights. In sans serif for example, 4,5 and 7 are smaller
> than the rest and for bold math only 7 is bigger than the rest.
>
> I would certainly be a little perplexed finding I would get different
> sized fences depending on whether I was talking about R^3 or R^4 and
> its the kind of thing which would prevent me from using it in production.
I also believe that's a bug. After all, in Computer Modern the heights
are made in a way that the brackets have the same size, and Latin Modern
is aimed at being compatible with CM, as far as I understand.

Moreover, even if that were a totally different font, the point raised
by Morten was exactly what happened in my text when I saw the strange
behaviour: I was saying that dim(R^{2x3}) is different from dim(R^4),
and saw the different sizes.