[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

revised report of EuroTeX math font BOF

Before I forget, here is the revised version of my report about the
math font BOF sessions at EuroTeX.  I've incorporated Barabara's and
Taco's sugesstions and added a few references to related information,
including the now official REC-MathML, which was released last week.

If you think that there is anything to be added or corrected, please
let me know.  Otherwise, you're welcome to go aheead and publish it
in TUGboat/Baskerville/MAPS/CahiersGUT/whatever.

Cheers, Ulrik.

P.S. The list of addresses primarily lists members of this group who
were present at the BOF sessions, but there were other participiants,
whose names I have left out.  Did anybody keep a list?  Barbara?

P.P.S. I tentatively included Joerg's new address (still commented
out), subject to confirmation and or correction.  What's the status?

%%% ====================================================================
%%%  @TeX-file{
%%%     author          = "Ulrik Vieth",
%%%     email           = "vieth@thphy.uni-duesseldorf.de",
%%%     filename        = "mfg-rep.tex",
%%%     version         = "0.04",
%%%     date            = "12 April 1998",
%%%     time            = "04:04:00 CEST",
%%%     checksum        = "27812 379 2241 16125",
%%%     codetable       = "ISO/ASCII",
%%%     keywords        = "EuroTeX 98, STIX, MathML, math symbol fonts",
%%%     supported       = "yes",
%%%     abstract        = "",
%%%     docstring       = "The checksum field above contains a CRC-16
%%%                        checksum as the first value, followed by the
%%%                        equivalent of the standard UNIX wc (word
%%%                        count) utility output of lines, words, and
%%%                        characters.  This is produced by Robert
%%%                        Solovay's checksum utility.",
%%%  }
%%% ====================================================================

%% v 0.01 -- 1998/04/02   initial version, sent to bb, Pattrick and Taco.
%% v 0.02 -- 1998/04/10   revised version, incorporated bb's corrections,
%%                        moved all URL references to bibliography.
%% v 0.03 -- 1998/04/12   update references for official MathML release.
%% v 0.04 -- 1998/04/12   more finishing touches

  author =       "Alan Jeffrey",
  title =        "{Math font encodings: A workshop summary}",
  journal =      "TUGboat",
  year =         "1993",
  volume =       "14",
  number =       "3",
  pages =        "293--295",
  author =       "Richard J. Kinch",
  title =        "{MetaFog: Converting {\MF} shapes to contours}",
  journal =      "TUGboat",
  year =         "1995",
  volume =       "16",
  number =       "3",
  pages =        "233--243",
  author =       "Matthias Clasen and Ulrik Vieth",
  title =        "{Towards a new math font encoding for {\AllTeX}}",
  journal =      "Cahiers GUTenberg",
  year =         "1998",
  volume =       "28--29",
  pages =        "94--121",
  note =         "Proceedings of the 10th European \TeX{} Conference",
  title =        {{Home Page of the Math Font Group (MFG)}},
  howpublished = {\url{http://www.tug.org/twg/mfg/}},
  note =         {Includes complete archives of test releases,
                  discussion papers and mailing list traffic.},
  author =       {\acro{STIPUB} Working Group},
  title =        {{\acro{STIX} symbol tables}},
  howpublished = {\url{http://www.ams.org/STIX/}},
  author =       {Unicode Consortium},
  title =        {{Unicode 2.0 symbol tables}},
  howpublished = {\url{http://www.unicode.org/Unicode.charts/}},
  author =       {W3C Math Working Group},
  title =        {{MathML symbol tables}},
  howpublished = {\url{http://www.w3.org/Math/mathmlchars.pdf}},
  author =       {W3C Math Working Group},
  title =        {{W3C Official Recommendation:
                  Mathematical Markup Language~1.0}},
  howpublished = {\url{http://www.w3.org/TR/REC-MathML/}},



\title{Summary of math font-related activities at~Euro\TeX~'98}

\def\rtitlex{MFG discussion document}
\def\midrtitle{{\sl Task 1: Organization\/}}



The subject of math symbol fonts has been one of the major topics of
interest at the 10th European \TeX{} Conference (Euro\TeX~'98), which
was held on March~29--31, 1998 at St.~Malo, France as part of the 2nd
Week on Electronic Publishing and Digital Typography

During the conference a paper summarizing the activities of the
Math Font Group%
\footnote{also known as: \LaTeX{}3 Project / TUG Technical Working
  Group on extended math font encodings (\hbox{WG 92-01})}
(\acro{MFG})~\cite{URL:MFG} was presented and two \acro{BOF} sessions
on math fonts were held, bringing together members of the \acro{MFG}
and representatives of other interested parties, such as the
\acro{W3C} MathML working group, the \acro{STIX} project, as well as
publishers and typesetters.

In addition, there were also many private discussions on math fonts at
lunches, dinners, and at informal get-togethers in the local caf{\'e}s or

%% bb -- the space around these items does stand out rather much
%%       relative to other vertical white space on the page; guess
%%       i should look into ltugboat.cls ...
The discussions at the \acro{BOF} sessions primarily revolved around
two major topics:
\item the organization of math symbols in general, including their
  representation on the \acro{WWW},
\item the development and implementation of new \hbox{8-bit} math
  fonts for \AllTeX{} in particular.

\section{Organization of math symbols}

On the first topic, Barbara Beeton and Patrick Ion of the \acro{AMS}
provided some information about the so-called \acro{STIX} project,
which is driven by a group of scientific and technical publishers

So far, the primary goal of the \acro{STIX} project has been to
compile a comprehensive list of \emph{all} math symbols used by the
participating publishers (also including what many people might call
``unreasonable'' ones), to document their intended meanings, and to
provide examples of their use in support of an application to the
Unicode Consortium and the \acro{ISO} working group on coding

A preliminary list of symbols has already been submitted to Unicode in
March~1998, but it appears that there are quite a few symbols that
have been missed, so the Unicode submission will have to be followed
up when more material is available.

Apart from compiling a comprehensive list of math symbols, there is
also a commitment to commission the production of a set of
high-quality fonts implementing all the symbols, which should be
freely distributable.  It is hoped that the availability of such a
font set will be a crucial step to help promote the use of MathML on
the World Wide Web without being restricted to the symbol complement
provided by the system fonts.

Information about the list of symbols collected by the \acro{STIX}
project currently resides on internal pages on the \acro{AMS} Web
server~\cite{URL:STIX} and is kept in a format similar to the Unicode
symbol tables~\cite{URL:Unicode}, but there are plans to release a
printable version of these tables in \acro{PDF} format to the general
public soon.

It was pointed out that a printable version of the symbol tables from
the MathML specification is supposed to be available in \acro{PDF}
format on the \acro{W3C} Web server~\cite{URL:MathML}.%
\footnote{Unfortunately, these symbol tables seem to have disappeared
  about a week after the conference when the MathML `Proposed
  Recommendation' was updated and promoted to an `Official Recommendation'.
  Hopefully, they will be put back after they have been updated as well.}
The latest version of the MathML specification also includes some
background information about the \acro{STIX} project and references to
various other glyph collections~\cite[Chapter~6]{URL:REC-MathML}.

\section{Implementation of new 8-bit math fonts for \protect\AllTeX}

On the second topic, the status of the activities of the Math Font
Group was reported in a conference paper~\cite{Vieth:Euro98} presented
by Ulrik Vieth in the morning session on the first day of the

So far, a set of encodings for new 8-bit math fonts for \AllTeX{} has
been developed based on a proposal dating back to \acro{TUG}~'93
\cite{Jeffrey:TB14-3-293}, which aims to fulfill certain design goals
and to satisfy a number of technical constraints.

These encodings, which consist of three primary encodings and several
additional ones, have been implemented as a set of virtual font using
glyphs taken from existing or newly-developed \MF{} or \PS{} fonts.
Several such sets of virtual fonts have been developed, covering most
of the presently available sets of math fonts usable with \TeX{}, but
the implementation unfortunately remains incomplete in several cases.
It also doesn't yet take into account many of the symbols identified
by the \acro{STIX} project, which may have to be added to the proposed
encodings if they are really needed.

A \LaTeX{} interface to access the new encodings and to switch between
different font sets implementing these encodings is already in place
and may be used either as a module to build a modified \LaTeX{} kernel
or as an add-on package for use with standard \LaTeX{}.  However, a
Plain \TeX{} interface is still missing and remains to be developed.

Given all these preparations, the question remains whether the
encodings developed by the \acro{MFG} are acceptable to the user
community and whether they satisfy the needs of scientific and
technical publishers.  While there wasn't a clear answer to this
question, there seemed to be a consensus that new 8-bit math fonts
addressing the organizational problems of the old 7-bit math fonts
were indeed needed and that the work of the \acro{MFG} provides a
suitable starting point, which may have to be refined later during the
implementation process.

In particular, there was a suggestion to relax the strict requirement
for compatibility with Plain \TeX{} or \LaTeX{} within the first four
math families, and to adopt a slightly more rational organization
which would allow to have fewer missing glyphs in some implementations
of the primary symbol font by leaving out some problematic glyphs and
relocating them to one of the extra symbol fonts.

Another request came from A.~Berdnikov, the coordinator of the 8-bit
Cyrillic encodings for \TeX{}, who pointed out that Russian math
typesetting traditions required different shapes of big operators,
such as upright integrals and bigger versions of summation and product
signs, and asked to support such variants in the new math font
encodings as well.

Since it is clear that it will be necessary to add several additional
symbol font encodings if all the \acro{STIX} glyphs are to be
incorporated eventually, minor adjustments to the present proposal may
be needed anyway and should not present a problem.  In any case, the
encoding tables presented at the conference should not be taken as the
final word.

A strong driving force to push forward the implementation of new 8-bit
math fonts for \AllTeX{} came from Taco Hoekwater of Kluwer Academic
Publishers.  As part of his professional activities, he is currently
working on a project to implement as many mathematical symbols as
possible in Type~1 format by the end of this year, possibly including
everything in the list of \acro{STIX} glyphs.

Since Kluwer Academic Publishers consider their products to be
journals and books, not fonts, Taco is allowed to put all the fonts he
produces for Kluwer into the public domain.  He has already converted
several existing \MF{} symbol fonts (including \texttt{rsfs},
\texttt{stmary} and \texttt{wasy}) to Type~1 format using the MetaFog
converter \cite{Kinch:TB16-3-233}, and released the results to
\acro{CTAN} shortly before the conference.

Concerning the production of new 8-bit math fonts, he suggested
concentrating on the Computer Modern version which appears to be the
easiest one to start with.  In particular, he proposed to start by
de-virtualizing the present implementation, which happens to draw
characters from a number of base fonts, so as to have a real \MF{}
font that could be converted more easily with MetaFog.

New symbols from the \acro{STIX} collection could then be added by new
\MF{} designs, which shouldn't be too difficult to develop in most
cases, as many symbols can be realized by combinations or variations
of existing symbols.

On the other hand, there seems to be little that can be done about the
versions based on commercial font sets such as MathTime or Lucida New
Math, which will probably remain restricted to whatever symbol
complement is provided in the present versions of the base fonts,
unless the suppliers of these font sets will invest some work

Another suggestion also discussed was to have a set of 8-bit fonts
serving as glyph containers organized by types of symbols, which could
either be used as the basis for a virtual font implementation of 8-bit
math fonts, meeting the technical constraints of \TeX{}, or combined
into a single huge 16-bit math font for Omega.  While this might be an
interesting option for the future, it was pointed out by several
participants that neither Omega nor virtual fonts could be assumed to
be available everywhere and that a straightforward \MF{}
implementation of new 8-bit fonts for \TeX{} was still needed.

Finally, it was discussed what to do about the Plain \TeX{} support of
the new math fonts.  Since Kluwer Academic Publishers are using Hans
Hagen's Con\TeX{}t package, which happens to be based on Plain \TeX{},
Taco will take care of this task as well, since he will need it for
his own work.

A suggestion to use the existing \LaTeX{} support on top of a Plain
\TeX{} emulation of the \acro{NFSS} interface was rejected, since the
\LaTeX{}-like syntax doesn't easily fit into the framework of the
Con\TeX{}t system, so a low-level Plain \TeX{} interface is preferred.

\section{Summary and Conclusions}

In summary, one might say that the Euro\TeX~'98 conference was a great
success for math font-related activities in that it helped to bring
together members of several working groups and other interested
parties, who so far have been working on closely related topics
independently of each other.

In particular, the \acro{STIX} project provided a lot of input to the
8-bit math font encodings for \TeX{}, while on the other hand there
was also some feedback to the \acro{STIX} project in the form of
additional symbols that have been missed so far.

While the \acro{STIX} project will continue to work on getting their
list of symbols assigned to Unicode, primarily in support of MathML
and \acro{SGML}-based authoring systems, the Math Font Group will
continue to work on completing the implementation of new 8-bit math
fonts for \AllTeX{}, hopefully by the end of this year.

Current plans include starting with the development of the Plain
\TeX{} interface and a de-virtualized \MF{} implementation of the
Computer Modern version as soon as possible and to begin adding more
symbols once this is in place.

It is hoped that we will be able to provide at least one very
comprehensive implementation of new 8-bit math fonts (including the
\acro{STIX} glyphs) in Computer Modern style in both \MF{} and Type~1
formats.  In addition, we weill provide several partial
implementations for other font families such as MathTime and Lucida
New Math, which will be implemented as virtual fonts based on the
symbol complement provided by these font sets.

%% anything else?


%% any other participants at the BOFs?

\author{Barbara Beeton}
\address{American Mathematical Society\\
  Providence, RI, USA}

\author{Thierry Bouche}
\address{Universit{\'e} Joseph Fourier\\
  Grenoble, France}

\author{Taco Hoekwater}
\address{Kluwer Academic Publishers\\
  Doordrecht, The Netherlands}

\author{Patrick Ion}
\address{American Mathematical Society\\
  Ann Arbor, Mich., USA}

\author{J{\"o}rg Knappen}
  Mainz, Germany}
%  Heidelberg, Germany}
%\netaddress{knappen@springer.de} ???

\author{Chris Rowley}
\address{Open University\\
  London, UK}

\author{Ulrik Vieth}
  D{\"u}sseldorf, Germany}