Math fonts

[VP]

Dear TeX users

My layout prefers two set of math fonts.

1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.

My example:

\documentclass{article}
\begin{document}

This is my test.

\section{This is one $sfmath$}

This is my $normal math$ test.

\begin{figure}
\caption{caption $sfmath$...}
\end{figure}

\end{document}

Please advise how to proceed this automately without using
\sffamily... in "Section/Table/Figure captions" place to place.

thanks
-vp-

[Marc van Dongen]

Post by VP
My layout prefers two set of math fonts.
1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.
\documentclass{article}
\begin{document}
This is my test.
\section{This is one $sfmath$}
This is my $normal math$ test.
\begin{figure}
\caption{caption $sfmath$...}
\end{figure}
\end{document}
Please advise how to proceed this automately without using
\sffamily... in "Section/Table/Figure captions" place to place.

I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

If you decide to accept the advice, not only will you (IMO) do your
readers a favour, but also will you save yourself time because you
won't have to implement the section/cation styles.

Regards,


Marc van Dongen

[Sebastian Schubert]

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I guess this depends a bit on the field. If it is clear that a sans
serif letter in a headline is not a tensor because there are no tensors
anyway in the document and you wouldn't even expect one, I think it is
ok to adjust the symbol.

In my case, I prefer to use sans serif for figures. In the caption, I
want to use the same font.

Sebastian

[Donald Arseneau]

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I really hate it when people apply this argument to bold
titles, as in the common document classes. Nobody is going
to think a bold letter in bold test refers to a vector or
whatever. Instead the normal-weight math appears light and
so stands out, but with no specific meaning.


Donald Arseneau ***@triumf.ca

[Marc van Dongen]

Post by Donald Arseneau

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I really hate it when people apply this argument to bold
titles, as in the common document classes.  Nobody is going
to think a bold letter in bold test refers to a vector or
whatever.  Instead the normal-weight math appears light and
so stands out, but with no specific meaning.

I _love_ it when people apply bold arguments like yours: I don't think
it holds:-)

We've all seen plenty of examples of books with a chapter/section
title that are typeset in a different style. Since the typeface of the
running text is pretty uniform, changing the style for the title
usually doesn't pose a problem because the shape of the letters
doesn't matter: there is no variation; an a is an a.

When it comes to math, the shapes of the letters _does_ matter; an $a$
may not be an $\mathbf{a}$. Let's assume captions are typeset by
making them bold. How can you tell the meaning of the letter a in
\section{$a$} and that of the letter a in \section{$\mathbf{a}$}?
(Answer, unless you know the title typesetting convention you cannot
tell the difference.)

Regards,


Marc van Dongen

[Sebastian Schubert]

Post by Marc van Dongen

Post by Donald Arseneau

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I really hate it when people apply this argument to bold
titles, as in the common document classes. Nobody is going
to think a bold letter in bold test refers to a vector or
whatever. Instead the normal-weight math appears light and
so stands out, but with no specific meaning.

I _love_ it when people apply bold arguments like yours: I don't think
it holds:-)

...

Post by Marc van Dongen
When it comes to math, the shapes of the letters _does_ matter; an $a$
may not be an $\mathbf{a}$. Let's assume captions are typeset by
making them bold. How can you tell the meaning of the letter a in
\section{$a$} and that of the letter a in \section{$\mathbf{a}$}?
(Answer, unless you know the title typesetting convention you cannot
tell the difference.)

While I think you are right, theoretically and when strictly following
ISO, in my experience this is normally no issue. I guess most people
perceive more the relative difference to the surrounding text than the
actual form of the letter. I would state (and I've no way to prove it)
that in general using $a$ in a bold headline is more distracting than to
just use a bold a instead. One has to think why it does look different
while normally, the bold letter in normal math looks different.

Of course, if your text relies heavily on the difference between bold
and normal math, and common senses does not tell you immediately what
the symbol means, then one has to be careful and, actually I would not
use the symbol at all in the headline if that's possible.


Sebastian

[Marc van Dongen]

Post by Sebastian Schubert
While I think you are right, theoretically and when strictly following
ISO, in my experience this is normally no issue. I guess most people
perceive more the relative difference to the surrounding text than the
actual form of the letter. I would state (and I've no way to prove it)
that in general using $a$ in a bold headline is more distracting than to
just use a bold a instead. One has to think why it does look different
while normally, the bold letter in normal math looks different.
Of course, if your text relies heavily on the difference between bold
and normal math, and common senses does not tell you immediately what
the symbol means, then one has to be careful and, actually I would not
use the symbol at all in the headline if that's possible.

I agree that an $a$ (math) in a sectional unit title looks bad in
Computer Modern. I don't know a solution to using math that looks good
in sectional unit titles. I'd avoid it when possible.

Regards,


Marc van Dongen

[Donald Arseneau]

Post by Marc van Dongen

Post by Donald Arseneau

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I really hate it when people apply this argument to bold
titles, as in the common document classes.  Nobody is going
to think a bold letter in bold test refers to a vector or
whatever.  Instead the normal-weight math appears light and
so stands out, but with no specific meaning.

I _love_ it when people apply bold arguments like yours: I don't think
it holds:-)
We've all seen plenty of examples of books with a chapter/section
title that are typeset in a different style. Since the typeface of the
running text is pretty uniform, changing the style for the title
usually doesn't pose a problem because the shape of the letters
doesn't matter: there is no variation; an a is an a.
When it comes to math, the shapes of the letters _does_ matter; an $a$
may not be an $\mathbf{a}$.

How did you get off on that tangent!? My argument "holds" but you don't
have a "grip" on it. I'm talking color -- ie blackness, thickness, weight
-- not shape. Your substitution of a roman "a" changes both the slope and
even the letter form (only certain letters change form for italic,
typically a,f,g).

Post by Marc van Dongen
Let's assume captions are typeset by
making them bold. How can you tell the meaning of the letter a in
\section{$a$} and that of the letter a in \section{$\mathbf{a}$}?

(captions by \section?)
I would have *liked* to assume that a well-designed document class
allows me to type \section{$a$} and get a bold math italic a; ie,
the section command uses \bfseries\boldmath for font selection.
The user should not have to set the math-font explicitly, not
even with a proper $\bm{a}$, let alone an erroneous \mathbf{a}.
Claims that math should be medium-weight in a bold section title
make as much sense as claiming the math font should also be normalsize.

Post by Marc van Dongen
(Answer, unless you know the title typesetting convention you cannot
tell the difference.)

"Difference" is the point. Bold weight is no absolute thing, but
is a difference from the surrounding norm. Sticking "normal" weight
math into bold text makes it *abnormal*, whereas bold math looks normal
in bold text.

The usual facile argument against bold math is: "but people will misinterpret
a bold math letter as a vector/tensor/whatever", which is made-up nonsense.
Nobody who sees \boldmath math in \textbf text will think the boldness
has special significance.

The alternative argument: "How will the reader be able to distinguish
meaningful bold math from the other bold math" is not stupid, but
doesn't hold up either. First and most sensible advice: don't put such
maths into section titles! They don't belong there. If you can't
restrain yourself, then that's what heavy/black/extra-bold fonts are
for. (Unfortunately, proper support for extra-bold \mathbf{} within
boldmath is lacking in the standard LaTeX document classes and CM
fonts.)


Donald Arseneau ***@triumf.ca

[Peter Flynn]

On 25/01/12 16:10, Donald Arseneau wrote:
[...]

Post by Donald Arseneau
I would have *liked* to assume that a well-designed document class
allows me to type \section{$a$} and get a bold math italic a; ie, the
section command uses \bfseries\boldmath for font selection. The user
should not have to set the math-font explicitly, not even with a
proper $\bm{a}$, let alone an erroneous \mathbf{a}. Claims that math
should be medium-weight in a bold section title make as much sense as
claiming the math font should also be normalsize.

The risk is that some people will assume that a math symbol which is
required to be bold in normal circumstances, should be bolder-than-bold
if the surrounding text is already bold. A bit like the cutting from the
Washington Post some years ago where the style for an article headline
was italic, so whoever was setting it made the embedded emphasis on one
word "even more italic" by sloping the text more towards the right :-)

Post by Donald Arseneau
First and most sensible advice: don't put such maths into section
titles! They don't belong there.

This the best advice anyone writing mathematical text could have. I have
seen writers wanting entire display equations in section titles.

///Peter

[Marc van Dongen]

Post by Marc van Dongen

Post by Donald Arseneau

Post by Marc van Dongen
I'd advise against using a different typefaces for math in captions/
captions and the running text. The same letter in a different typeface/
type style usually has a different meaning in math. That's why \emph{$a
$} looks the same as $a$.

I really hate it when people apply this argument to bold
titles, as in the common document classes.  Nobody is going
to think a bold letter in bold test refers to a vector or
whatever.  Instead the normal-weight math appears light and
so stands out, but with no specific meaning.

I _love_ it when people apply bold arguments like yours: I don't think
it holds:-)
We've all seen plenty of examples of books with a chapter/section
title that are typeset in a different style. Since the typeface of the
running text is pretty uniform, changing the style for the title
usually doesn't pose a problem because the shape of the letters
doesn't matter: there is no variation; an a is an a.
When it comes to math, the shapes of the letters _does_ matter; an $a$
may not be an $\mathbf{a}$.

How did you get off on that tangent!?  My argument "holds" but you don't
have a "grip" on it.  I'm talking color -- ie blackness, thickness, weight
-- not shape.  Your substitution of a roman "a" changes both the slope and
even the letter form (only certain letters change form for italic,
typically a,f,g).

Post by Marc van Dongen
Let's assume captions are typeset by
making them bold. How can you tell the meaning of the letter a in
\section{$a$} and that of the letter a in \section{$\mathbf{a}$}?

(captions by \section?)
I would have *liked* to assume that a well-designed document class
allows me to type \section{$a$} and get a bold math italic a; ie,
the section command uses \bfseries\boldmath for font selection.
The user should not have to set the math-font explicitly, not
even with a proper $\bm{a}$, let alone an erroneous \mathbf{a}.
Claims that math should be medium-weight in a bold section title
make as much sense as claiming the math font should also be normalsize.

Post by Marc van Dongen
(Answer, unless you know the title typesetting convention you cannot
tell the difference.)

"Difference" is the point.  Bold weight is no absolute thing, but
is a difference from the surrounding norm.  Sticking "normal" weight
math into bold text makes it *abnormal*, whereas bold math looks normal
in bold text.
The usual facile argument against bold math is: "but people will misinterpret
a bold math letter as a vector/tensor/whatever", which is made-up nonsense.
Nobody who sees \boldmath math in \textbf text will think the boldness
has special significance.
The alternative argument: "How will the reader be able to distinguish
meaningful bold math from the other bold math" is not stupid, but
doesn't hold up either.  First and most sensible advice: don't put such
maths into section titles!  They don't belong there.  If you can't
restrain yourself, then that's what heavy/black/extra-bold fonts are
for.  (Unfortunately, proper support for extra-bold \mathbf{} within
boldmath is lacking in the standard LaTeX document classes and CM
fonts.)

Donald, I'm only using _your_ argument about boldness to show it
doesn't work. There's no convention about how bold a normal size
letter should be in a title to qualify for the title equivalent of the
letter.... If this cannot work for boldness factors, how is tis
supposed to work for different shapes.

Your argument about difference doesn't hold because the reader may not
have had the chance to compare the different versions of math in a
title: if only one version is used in the tile, then which one is
it???? The same holds for a letter that is typeset in sans serif "a"
and a math roman "a" in the running text. What if you see a sans serif
"a" in the section title? What symbol does it denote???? (I'm using
the sans serif example, because that's what triggered "it" all and to
show that boldness is irrelevant.)

I agree entirely with your suggestion that one shouldn't put _such_
math into section titles. But I don't buy the argument that (at least
in this case) people shouldn't do it because "it" doesn't work for
bold math. In my opinion, the argument should be that you shouldn't do
it because we have no accepted conventions of how to typeset math in
section titles. That's all.

Finally, I agree with your comment about well designed classes with
proper typefaces, but in reality, we don't have the luxury of multiple
shades of boldness, sans seriffeness, and so on. Most typefaces don't
support them....

Regards,


Marc van Dongen

[William F Hammond]

Post by Marc van Dongen
...
I agree entirely with your suggestion that one shouldn't put _such_
math into section titles. But I don't buy the argument that (at least
in this case) people shouldn't do it because "it" doesn't work for
bold math. In my opinion, the argument should be that you shouldn't do
it because we have no accepted conventions of how to typeset math in
section titles. That's all.

Absolutely.

Beyond the absence of such conventions is the situation with online
versions of LaTeX documents. A section title, particularly a first-order
section title, comes to be regarded as meta-date and thereby becomes a
candidate for processing by robots incapable of dealing with math.
Moreover, article section titles are imported as meta-data as a means
of conveying the contents of an article for casual scanning, and in
that context mathematical markup, even when properly rendered, is usually
not articulate.

Even if an author does not understand this, an editor should.

-- Bill

[Ilya Zakharevich]

Post by William F Hammond

Post by Marc van Dongen
...
I agree entirely with your suggestion that one shouldn't put _such_
math into section titles. But I don't buy the argument that (at least
in this case) people shouldn't do it because "it" doesn't work for
bold math. In my opinion, the argument should be that you shouldn't do
it because we have no accepted conventions of how to typeset math in
section titles. That's all.

Absolutely.

All that I may say is that I'm completely stunned. Are we discussing
haiku collections, or books on political science? (La)TeX has more
applications that this.

Let me recall that (according to Littlewood) Kronecker (?) said that a
"quality" of mathematicians should be measured by the quantity of bad
texts they have written. The work of initial discoverers is very hard
to read. There is a certain flavor of "major" math book which are,
essentially, incomprehencible (without a substantial mental effort)
since their purpose is to completely change the reader's existing
point of view on phenomena.

Usually, in 20-30 years, an alternative more user-friendly approach
might be found. On the other hand, in that time the new paradigms
might have "been percolating through osmosis", so it is not only an
effect of "better text", but also of "more prepared readers". Anyway,
reading such books is an unalienable part of life of mathematicians.

For most mathematicians, a significant part of their life is spent
trying to fight their way through such life-changing excercises.
(AFAIK, in all other vocations, the rate of such events is at least an
order of magnitude smaller.) And many "lesser" texts are written in
similar styles, because using "soft glove" approach, while feasible,
would bloat the text, so would make the understanding yet harder.
Having a semester-long group effort to work through a short paper is
not unheard of. (And, often, these papers are extremely
well-written.)

Ergo: in a book on math (as opposing to a book USING math) the primary
objective is to provide as much help to the reader as possible.
Following typographical conventions (most of which were honed on books
in different subjects!) is a distant second goal. And having PROPER
sectioning with APPROPRIATE names is one of the major ways to simplify
comprehention without bloating the volume. (Both of a way to
comprehend the large-block architecture of the arguments, AND a way to
navigate starting from the ToC.) For a large proportion of headings,
this means using math in names. And in parts of math where bold is
very significant, this may mean using bold in math in header names.

Post by William F Hammond
Beyond the absence of such conventions is the situation with online
versions of LaTeX documents. A section title, particularly a first-order
section title, comes to be regarded as meta-date and thereby becomes a
candidate for processing by robots incapable of dealing with math.
Moreover, article section titles are imported as meta-data as a means
of conveying the contents of an article for casual scanning, and in
that context mathematical markup, even when properly rendered, is usually
not articulate.
Even if an author does not understand this, an editor should.

This argument looks as wild to me as saying: since people may make
spelling mistakes when they write, all communication should be
restricted to oral form.

Yours,
Ilya

[Lee Rudolph]

Ilya Zakharevich <nospam-***@ilyaz.org> appears to quote
(who? please tell!)

Post by Ilya Zakharevich
Let me recall that (according to Littlewood) Kronecker (?) said that a
"quality" of mathematicians should be measured by the quantity of bad
texts they have written. The work of initial discoverers is very hard
to read. There is a certain flavor of "major" math book which are,
essentially, incomprehencible (without a substantial mental effort)
since their purpose is to completely change the reader's existing
point of view on phenomena.
Usually, in 20-30 years, an alternative more user-friendly approach
might be found. On the other hand, in that time the new paradigms
might have "been percolating through osmosis", so it is not only an
effect of "better text", but also of "more prepared readers".

...

Or, as the complex analyst Robert C. Gunning put it once, when someone
asked him, of a certain result he was quoting without proof, "Isn't
that deep?" (with the implication that at least a proof sketch would
be nice to have), "Mathematics rises to the surface with time."

_A propos_ of the actual subject of this thread, Gunning (and his
co-author Hugo Rossi) were, as far as I have been able to determine,
the first to use a "blackboard bold" font in a proper textbook
(namely, _Analytic Functions of Several Complex Variables_,
Prentice-Hall Series in Modern Analysis, 1965), and they *did*
use it in section titles, eg. Chapter IV, Section C, "Analytic
Sheaves on Subdomains of $\mathbb{C}^n$. (Curiously--I have just
noticed this now--in the body of the book this section heading
appears in bold roman, except for the \mathbb{C}, which is the
same very light roman that is used for the complex numbers in
the text; but in the table of contents [on pp. xi-xii, by the
way, preceded only by a dedication and three pages of preface;
definitely *not* "in the middle" of the book], the section
heading is in non-bold roman, and the \mathbb{C} is very light
*italic*.) How they persuaded Prentice-Hall to go along with
this, Gunning could not recall when I finally got around to
asking him several years ago. Of course a make-shift blackboard
bold had been used by the Princeton University Press earlier,
in their Mathematical Notes series, during the period when
the books were prepared on typewriters (pre-Selectric).

Lee Rudolph

[Ilya Zakharevich]

Post by Lee Rudolph
(who? please tell!)

Sorry for confusing layout: it was indented, but it is not a quote,
just something not DIRECTLY related to the topics of typesetting headers...

Ilya

[Sebastian Schubert]

Post by VP
Dear TeX users
My layout prefers two set of math fonts.
1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.
\documentclass{article}
\begin{document}
This is my test.
\section{This is one $sfmath$}
This is my $normal math$ test.
\begin{figure}
\caption{caption $sfmath$...}
\end{figure}
\end{document}
Please advise how to proceed this automately without using
\sffamily... in "Section/Table/Figure captions" place to place.

First, define a new mathversion (eg sans) with
\DeclareMathVersion{sans}
then set which letters you want to use with
\SetSymbolFont{letters}{sans}{OML}{..}{..}{..}
which depends on the font you want to use (look in the appropriate LaTeX
package. You can switch in the document with \mathversion{sans}. This
allows to use at least different latin letters in math. A complete
second math setup (with greek and symbols) is not easy to set up. I
tried it in the mdsymbol package with the onlysansmath option but with
it you get new sans serif symbols and have to still set up the main sans
serif font. (if you don't want to use the Adobe MyriadPro which I
adapted a package for)

Second, eg the koma classes allow to customize the fonts for different
parts of the document. In your case, something like
\addtokomafont{caption}{\sffamily\mathversion{sans}}
would do the trick.

HTH
Sebastian

[Guenter Milde]

Post by Sebastian Schubert

Post by VP
My layout prefers two set of math fonts.
1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.

...

Post by Sebastian Schubert
First, define a new mathversion (eg sans) with
\DeclareMathVersion{sans}
then set which letters you want to use with
\SetSymbolFont{letters}{sans}{OML}{..}{..}{..}

An example for a complete "sans" math version using cmbr fonts is given in
http://newsgroups.derkeiler.com/Archive/Comp/comp.text.tex/2007-09/msg00181.html

A (German) list sans-serif TeX math fonts is available at
http://milde.users.sourceforge.net/Matheschriften/matheschriften.html#sans-serif-grotesk

Missing in this summary is the "kpfonts" package
which provides a "sf" math version out of the box:

\documentclass{minimal}
\usepackage{fixltx2e}
\usepackage[T1]{fontenc}
\usepackage{kpfonts}
\usepackage{lmodern}

\begin{document}

Serif $\sin^2(a+b)$

and \sffamily \mathversion{sf} sans serif $\sin^2(a+b)$

\end{document}

Post by Sebastian Schubert
Second, eg the koma classes allow to customize the fonts for different
parts of the document. In your case, something like
\addtokomafont{caption}{\sffamily\mathversion{sans}}
would do the trick.

Günter

[Donald Arseneau]

Post by Guenter Milde

Post by Sebastian Schubert

Post by VP
My layout prefers two set of math fonts.
1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.

First, define a new mathversion (eg sans) with

Ugh.

Post by Guenter Milde
A (German) list sans-serif TeX math fonts is available at
http://milde.users.sourceforge.net/Matheschriften/matheschriften.html#sans-serif-grotesk
Missing in this summary is the "kpfonts" package

See also the packages sansmath and sfmath.



Donald Arseneau ***@triumf.ca

[Guenter Milde]

Post by Donald Arseneau

Post by Guenter Milde

Post by Sebastian Schubert

Post by VP
My layout prefers two set of math fonts.
1. Normal Times font in Text/para.
2. Sansserif font in Section/Table/Figure captions.

First, define a new mathversion (eg sans) with

A (German) list sans-serif TeX math fonts is available at
http://milde.users.sourceforge.net/Matheschriften/matheschriften.html#sans-serif-grotesk
Missing in this summary is the "kpfonts" package

See also the packages sansmath and sfmath.

Thanks for the pointers.

sansmath is an out-of-the-box workaround using text fonts for math with some
limitations:

No change to non-alphanumeic symbols and Greek letters. The
eulergreek/EULERGREEK options are suboptimal - resulting in upright small
Greek and "half-serif" capital Greek letters.
It cannot be used together with fixmath or isomath.

sfmath does not provide a "sans" math version, instead:

After including the package sfmath.sty, all maths is displayed with
sans serif fonts; there is no way to switch back to the original
behavior.

-- http://dtrx.de/od/tex/sfmath.html

Günter