MathGroup Archive 2004

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

Search the Archive

Publicon LaTeX conversion problems

  • To: mathgroup at
  • Subject: [mg51894] Publicon LaTeX conversion problems
  • From: blazer at (Mat Marcus)
  • Date: Thu, 4 Nov 2004 01:50:44 -0500 (EST)
  • Sender: owner-wri-mathgroup at


I am using Mathematica and Publicon to write a book. The book is up to
about 100 pages and growing quickly. The final artifact needs to be in
AMS LaTeX. Once or twice each week I "Save as LaTeX...", run the
output through pdflatex and send it to the "beta testers"/reviewers. I
have encountered a number of issues and I am hoping to learn better
ways of coping with them by posting here.

My first question is simple. I need to use some symbols that the tex
conversion process does not directly support. In particular I need to
use \[Implies],\[DoubleLeftRightArrow],\[NotVerticalBar],\[NotCongruent],\[RightUpVector]
and others. I would like these to appear as \Longrightarrow,
\iff,\nmid,\nequiv,\upharpoonright respectively. Instead they are
emitted as the non-existent tex commands: (respectively)
\wolframcharacter{Implies}, \unicode{21d4}, \unicode{2224},
\unicode{2262}, etc. I have tried several workarounds within Publicon.

First I tried to make a special style that would allow raw TeX
commands to be emitted unchanged. I would have liked to replace the
"Automatic" conversion routine in the TeX ConversionRules with
something like an identity function, but it doesn't seem like one
exists. I tried InputFormToString and also StripFirstWord["Number"] in
the hopes of simulating the identity function, but the conversion
process seeems to stumble when these functions the corresponding
styles are inlined in an equation cell.

I tried creating a special style for each symbol that would emit the
appropriate character, but I found this to be too cumbersome.

I tried openinig the notebook (with embedded private style sheet) in
Mathematica 5.x, but Mathematica crashed when saving it as LaTeX.

I tried grabbing the sources for the LaTeX conversion from the web and
dropping them into the update folder for Publicon. The hope was to
possibly add the "missing" characters to the appropriate table in the
conversion source files. Unfortunately I couldn't figure out the
appropriate changes to the inti.m file to get the new sources to be

I tried contacting Premiere support, but haven't received useful
answers yet.

So now I post process with a sed script but I would like to find a
better solution entirely within Publicon.

My second issue is related. I have some source code snippets, and I
would like to associate some conversion rules to a style that would
suppress the escaping of backslashes, ampersands, etc. Then I could
wrap all of the emitted source code in a \lstlisting block. An
"identity"/no-escape conversion function would seem to help here as

Third issue: graphics are always output as .eps files. UNfortunately,
pdflatex cannot handle these. Is it possible to get the graphics to be
output in other formats? Or might it help to embed the graphics in the
latex instead? It seems like there is some provision for this in the
source code for LaTeX conversion, but I can't figure out how to
SetOptions to the right state in Publicon.

Any advice?


Mat Marcus
Senior Computer Scientist
Adobe Systems, Inc.

  • Prev by Date: Re: fractal image compression
  • Next by Date: Re: LegendreP (Symbolic) is different in Mathematica5 than previous versions (M4, M3 ..)
  • Previous by thread: Re: LegendreP (Symbolic) is different in Mathematica5 than previous versions (M4, M3 ..)
  • Next by thread: Re: Publicon LaTeX conversion problems