[XeTeX] XeLaTeX and math --- a structural design query

Ross Moore ross at ics.mq.edu.au
Thu Jun 18 07:36:06 CEST 2009

Hello Nicholas,

On 18/06/2009, at 1:34 PM, Nicolas Vaughan wrote:

> Hello,
> I have been commissioned by a university to design a collection of  
> monographs on mathematics. I should be providing a class file, a  
> user manual and a typeset example.
> At the outset, I decided to work in XeLaTeX, since this is the  
> system I have been working with lately, and with which I feel quite  
> comfortable.

The class file that you provide for others does not need to be
the same as what you end up using yourself.

It is my belief that the editor should not be forcing authors
to use packages and/or software tools with which they may
be unfamiliar or uncomfortable working with.
Authors need to put their knowledge into a compuscript
(e.g., LaTeX source) in as accurate and faithful a way
as is possible for *them* to do. Anything that is restrictive
or prescriptive can easily hamper this process, and lead to
a loss of quality or value of the content presented within the
final document.

> However, some folks have told me that, since I know not what the  
> precise content of the monographs will be (apart from the usual  
> LaTeX packages and AMS commands and environments), there is a risk  
> there might be a clash between XeLaTeX and the content (e.g., with  
> commutative diagrams, perhaps?).
> What do you suggest? Should I stick to plain old LaTeX, or should I  
> go all the way through with XeLaTeX?

You could develop slightly different class files according
to the different typesetting engines that authors might use.

The class files should contain as little as possible, concentrating
just on things that are truly helpful (e.g., paper-size, typeset area
on the page, settings for heading sizes and spacing, bibliography style,
equation numbering style, etc.
It could automatically load some standard packages that you expect
they'll probably use anyway; (e.g., AMSmath, array, dcolumn, hyperref,
natbib) especially if these help in setting-up the preferred styles
for headings, citations, etc.

Beyond this, the authors should be free to use whatever extra packages
they find convenient, and use LaTeX, pdf-LaTeX or XeTeX, according to
what they are used to using.
(Some mathematicians I know still even use AMS-TeX or Plain-TeX.)

It is then your job as editor to combine the submissions into
a volume that looks coherent and presents in a consistent way.
But it would be quite wrong for you to be imposing your own
preferences upon your authors, for the sake of your own convenience.
To me, this would have a serious risk of being counter-productive.

If some authors provide graphics that are of low resolution,
poorly labelled, or just look plain clunky, then you can rework them
to obtain a better quality. I've had to do this (quite gladly) with
every one of the Proceedings-like volumes and monographs that I've
helped to produce, over the years.

As for some more practical advice, the best package for the layout
of commutative diagrams is almost certainly Xy-pic --- but to get
the best results from it you need to use its PostScript back-end.
This can be done with pdfLaTeX and XeLaTeX only if you use special
packages like pdfTricks, or eps4pdf (and others), to create separate
graphics to be imported back into the main job.
This is not something that you can expect most authors to be able
to cope with. Similar comments apply to use of PS-tricks.

But restricting to using just ordinary (non-pdf) LaTeX loses out
in other ways. Similarly forbidding XeTeX cannot be a good idea
either, as someone may well want this for special fonts.

> Thanks for any ideas.
> Cheers!
> Nicolas Vaughan

Hope this helps,


Ross Moore                                       ross at maths.mq.edu.au
Mathematics Department                           office: E7A-419
Macquarie University                             tel: +61 (0)2 9850 8955
Sydney, Australia  2109                          fax: +61 (0)2 9850 8114

More information about the XeTeX mailing list