spacing of mathrm

For feedback and discussion for testers of the MathTime Professional II fonts.

Moderators: PTIForAdmin, WaS, Michael Spivak

zedler
Posts: 15
Joined: Fri Mar 03, 2006 1:50 am

spacing of mathrm

Post by zedler »

Hello,

\documentclass{book}
\usepackage{times,mtpro2}
\begin{document}
$(\mathrm j$
\end{document}

gives touching glyphs.

Michael
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: spacing of mathrm

Post by Michael Spivak »

zedler wrote:Hello,

\documentclass{book}
\usepackage{times,mtpro2}
\begin{document}
$(\mathrm j$
\end{document}

gives touching glyphs.

Michael
There's not much I can do about that---if you are using Times as the
text font, then in text (j also touches! [though $(\mathrm j$ is worse, with more overlap].

I'm wondering how this arose. Assuming that you didn't really want
\mathrm{(j ... I would guess that you are using roman letters as a set
of variables, either in addition to, or in place of, the MTPro2 italic letters.

In that case, you really would want a special font for this purpose, in the
same way that MTPro's \mathbf font has different spacing than the
Times-bold, so that subscripts and superscripts will work better.
zedler
Posts: 15
Joined: Fri Mar 03, 2006 1:50 am

Re: spacing of mathrm

Post by zedler »

There's not much I can do about that---if you are using Times as the
text font, then in text (j also touches! [though $(\mathrm j$ is worse, with more overlap].

I'm wondering how this arose. Assuming that you didn't really want
\mathrm{(j ... I would guess that you are using roman letters as a set
of variables, either in addition to, or in place of, the MTPro2 italic letters.

In that case, you really would want a special font for this purpose, in the
same way that MTPro's \mathbf font has different spacing than the
Times-bold, so that subscripts and superscripts will work better.
Yes, I really want to typeset $\exp(\mathrm j\omega\tau=$ ;-)
I suppose this can only be corrected by increasing the bracket side bearings, but your approach was to have very tight bracket side bearings and adjust/increase the spacing using kerns. This of course fails for \mathrm...
The tight bracket side bearings were also an issue in my previous example, $\[\]_{xy}$. CM, Fourier, Lucida and MnSymbol don't have this problem...

Michael
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: spacing of mathrm

Post by Michael Spivak »

zedler wrote:
There's not much I can do about that---if you are using Times as the
text font, then in text (j also touches! [though $(\mathrm j$ is worse, with more overlap].

I'm wondering how this arose. Assuming that you didn't really want
\mathrm{(j ... I would guess that you are using roman letters as a set
of variables, either in addition to, or in place of, the MTPro2 italic letters.

In that case, you really would want a special font for this purpose, in the
same way that MTPro's \mathbf font has different spacing than the
Times-bold, so that subscripts and superscripts will work better.
Yes, I really want to typeset $\exp(\mathrm j\omega\tau=$ ;-)
I suppose this can only be corrected by increasing the bracket side bearings, but your approach was to have very tight bracket side bearings and adjust/increase the spacing using kerns. This of course fails for \mathrm...
The tight bracket side bearings were also an issue in my previous example, $\[\]_{xy}$. CM, Fourier, Lucida and MnSymbol don't have this problem...

Michael
Actually, I didn't, and one can't, adjust the spacing after a left parentheses or before a right parenthesis
using kerns ;
even if you put kerns into the tfm file, they are ignored because the left parenthesis is an "opening", which determines its own spacing, and similarly the right parenthesis is a "closing". I chose side bearings for
the parenthesis that worked well with the italic letters on the math italic font.

Even if that were not the case, the real problem is that
in the expression \exp(\mathrm j the ( comes from the math italic font,
while the j is coming from an entirely different font, the Times-Roman font, and TeX has no way of kerning characters in different fonts. If you
were to use some other roman font as your text font, then the problem might very well be less or much more---it would depend entirely on the left side bearing of j in that particular font.

I suspect that j is being used here as a some special character (perhaps in electrical engineering, although I thought they preferred bold j); in that case, I would just define something like \myj to give a small kern followed by j---in fact, it's easier to type \myj than to type \mathrm j.

Sorry that [] doesn't work out for you, but I've never seen something like that in any mathematics paper, and since I like the way brackets work with the math italic characters in general, I wouldn't want to change the side bearings just for this special case (once again, changes couldn't be
overridden with kerns).
zedler
Posts: 15
Joined: Fri Mar 03, 2006 1:50 am

Re: spacing of mathrm

Post by zedler »

Sorry that [] doesn't work out for you, but I've never seen something like that in any mathematics paper, and since I like the way brackets work with the math italic characters in general, I wouldn't want to change the side bearings just for this special case (once again, changes couldn't be
overridden with kerns).
I can apply manual spacings, the "\mathrm j" is stored in a macro anyway and the empty brackets I need only once. Perhaps you're interested, I've put together a collection showing how different math font setups behave in the above mentioned cases: http://www.hft.ei.tum.de/mz/mtpro2_sidebearings.pdf

Michael
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: spacing of mathrm

Post by Michael Spivak »

="zedler
I can apply manual spacings, the "\mathrm j" is stored in a macro anyway and the empty brackets I need only once. Perhaps you're interested, I've put together a collection showing how different math font setups behave in the above mentioned cases: http://www.hft.ei.tum.de/mz/mtpro2_sidebearings.pdf

Michael
Interesting.

I'd say that CM looks the worst (especially the \omega and
\tau, as well as being so thin).

Lucida is somewhat "klunky", though definitely easy to read!
(Is this Lucida or Lucida-Bright?) Some one mentioned that section headings are sometimes printed in sans-serif,
so that a sans-serif math might be nice to have; I suspect that the Lucida
greek letters would work well for that.

If \mathrm j is in a macro, then probably there should also be some space on the right; certainly needed for CM, not really needed for Lucida
or Minion, useful for Fourier and MTPro2.

By the way, what is []_{\langle6\times6\rangle} ?
zedler
Posts: 15
Joined: Fri Mar 03, 2006 1:50 am

Re: spacing of mathrm

Post by zedler »

(Is this Lucida or Lucida-Bright?)
Pctex's Lucida fonts.
By the way, what is []_{\langle6\times6\rangle} ?
Excerpt from a paper (Deadline tomorrow Mar 30, Hawai time ;-))

Code: Select all

\begin{document}\let\mathbf\mbf
...
Next, the impedance matrix of the outer 12-port is obtained by
inverting $\mathbf{Y}^{\langle 16\times16\rangle}$ and taking the
upper left $\langle 12\times12\rangle$ submatrix
\begin{equation}
  \mathbf{Z}^{\langle 12\times12\rangle}=\left[{\mathbf{Y}^{\langle 16\times16\rangle}}^{-1}\right]_{\langle 12\times12\rangle}
\end{equation}
where the operator $[\,]_{\langle 12\times12\rangle}$ denotes taking
the submatrix. The $\mathbf Z^{\langle 6\times 6\rangle}=\mathbf Z$
matrix of the outer six-port is obtained by
Perhaps not the best notation, do you have a better idea? BTW, quite funny that both you and my boss are aficionados of differential forms ;-)

Wish you wedge and hodge,
Michael
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: spacing of mathrm

Post by Michael Spivak »

zedler wrote: Excerpt from a paper (Deadline tomorrow Mar 30, Hawai time ;-))

Code: Select all

\begin{document}\let\mathbf\mbf
...
Next, the impedance matrix of the outer 12-port is obtained by
inverting $\mathbf{Y}^{\langle 16\times16\rangle}$ and taking the
upper left $\langle 12\times12\rangle$ submatrix
\begin{equation}
  \mathbf{Z}^{\langle 12\times12\rangle}=\left[{\mathbf{Y}^{\langle 16\times16\rangle}}^{-1}\right]_{\langle 12\times12\rangle}
\end{equation}
where the operator $[\,]_{\langle 12\times12\rangle}$ denotes taking
the submatrix. The $\mathbf Z^{\langle 6\times 6\rangle}=\mathbf Z$
matrix of the outer six-port is obtained by
Perhaps not the best notation, do you have a better idea? BTW, quite funny that both you and my boss are aficionados of differential forms ;-)

Wish you wedge and hodge,
Michael
Not really, but I would probably have used something like UL_{\langle
12\times\12\rangle}(...) with U and L roman (or perhaps bold). And I
probably would actually have used something like UL_{[12]}, with
the idea that for square matrices [12] would mean \langle12\times12\rangle.
Locked