[Date Index] [Thread Index] [Author Index]

On typesetting

• To: mathgroup at smc.vnet.net
• Subject: [mg76605] On typesetting
• From: Selwyn Hollis <sh2.7183 at earthlink.net>
• Date: Thu, 24 May 2007 06:19:18 -0400 (EDT)
• References: <f2u57d$k2n$1@smc.vnet.net> <200705230940.FAA24349@smc.vnet.net>

One of the things I had hoped for in Mathematica 6 was improvement in  
its typesetting. I admit that I'm probably more picky about this sort  
of thing than most, but beautifully typeset mathematics is so common  
nowadays that sloppy typesetting really sticks out like a sore thumb  
to someone with an eye for such things. Now of course people have  
different tastes and preferences for some aspects of spacing,  
heights, script sizes, and so on, and Mathematica does allow user  
control of much of that, but there is one particular aspect over  
which the only control is a lot of tedious tweaking. Basically,  
AutoSpacing does a very poor job with multiplication spacing.  
Multiplication spaces are generally too wide, and on-screen they are  
horribly inconsistent. Apparently, loose spacing is preferred at WRI  
(which may be related to TeXForm's insistence upon putting \, spaces  
everywhere), and that's fine. As best I can tell, a multiplication  
space usually amounts to a \[ThinSpace], or 3/18 em. I find that much  
space horribly loose in most cases. A \[VeryThinSpace], or 1/18 em,  
is usually too thin, but often preferable. To my taste, 2/18 em would  
probably be a good compromise.

I've discovered that a ZeroWidthTimes option has been introduced with  
6.0, so apparently someone at WRI is thinking about these things. But  
ZeroWidthTimes->True only causes the exact opposite problem. Why not  
a TimesWidth option with "Tight", "Normal", and "Loose" as possible  
values?

In addition, Mathematica 6 has introduced two rather egregious new  
typesetting "issues:"

(1) Embedding equations in text ruins line-spacing. There is excess  
space below any line that contains an equation, even if the equation  
is as simple as y=x.

(2) In fractions set with ScriptLevel->0, the distance between the  
bar and the numerator's baseline has ballooned to almost the x-height  
of the font.

I have prepared a brief pdf document showing various comparisons that  
I hope make all this clearer. See it here:
http://www.math.armstrong.edu/faculty/hollis/typesettingwoes.pdf


- Selwyn Hollis

• Follow-Ups:
  • Re: On typesetting
    • From: Murray Eisenberg <murray@math.umass.edu>

• References:
  • Re: Evaluation Question
    • From: siewsk@bp.com