MathGroup Archive 2000

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

Search the Archive

Re: TeXSave undefined control sequences

  • To: mathgroup at
  • Subject: [mg26271] Re: TeXSave undefined control sequences
  • From: paulh at (P.J. Hinton)
  • Date: Sun, 10 Dec 2000 00:19:43 -0500 (EST)
  • Organization: Wolfram Research, Inc.
  • References: <90kprf$>
  • Sender: owner-wri-mathgroup at

On 6 Dec 2000 02:28:47 -0500, "Stephen P Luttrell"
<luttrell at> wrote:

> I have version "4.0 for Microsoft Windows (July 26, 1999)".

> 1. There are some undefined control sequences, such as \AlignmentMarker and

\AlignmentMarker is generated as the result of converting the
Mathematica character with longname \[AlignmentMarker].  These
characters do not map directly to any TeX control sequence, so you
could define the sequence to expand to \relax.

> \MathBegin{MathArray}...\MathEnd{MathArray}. 

These sequences are defined in notebook2e.sty.  Are you sure these
files are in a location that is searched by your TeX installation?
Did you remember to update the filename search database that is used
by your implementation of TeX?

> I use alignment markers in Mathematica a lot to get my equations 
> formatted nicely, so I need this facility to carry across to TeX cleanly.

As I mentioned above, there is no clean analog of \[AlignmentMarker]
in TeX.  In LaTeX 2e, one may use the eqnarray environment to control
layout of equations.  The closest thing that you can do in a
Mathematica notebook is to use grid box structures.  Paste this Cell[]
expression into a notebook and click on "Yes" on the resulting dialog
box to see what I am talking about.

          RowBox[{"a", "+", "b"}], "=", "c"},
          RowBox[{"c", "+", "d"}], "=", "e"}
        }]], "Input"]

This will convert to:

	a+b&=&c \\

> 2. Automatic numbers are not handled correctly. They come across as numbers
> that increment incorrectly.

The notebook that is converted by TeXSave[] is actually a preprocessed
version supplied by the front end; automatic numbering objects
(CounterBox[]) are resolved to literal values.  This is akin to
performing the front end menu command Input -> Convert Automatic
Objects to Literal before invoking TeXSave[].  Unfortunately, the
front end does not properly resolve the value of options with
CounterBox[] objects.  This hits automatically numbered equation
styles, which embed CounterBox[] objects in the value of the option

> 3. There was a transient problem when generating the MathBold5 fonts
> (something about floating point overflow), but this error occurred only
> once.

That might have to do with some troubles that DVIPS has with partially
subsetting the Type 1 Math fonts.  If you run into this problem, try
including the command line flag -j0 to force complete embedding of the
fonts.  This will result in larger PostScript output files, but the
glyphs in the file will be rendered properly on any PostScript
interpreting device.

P.J. Hinton
User Interface Programmer                         paulh at
Wolfram Research, Inc.

  • Prev by Date: Re: Big trouble with Mathematica Code parsing -- Rant.
  • Next by Date: Re: How to use the multiple varible Interpolation?
  • Previous by thread: TeXSave undefined control sequences
  • Next by thread: Re: TeXSave undefined control sequences