[XeTeX] fancyhdr and plyglossia

[Ross Moore]

Hi Vafa,

On 31/01/2010, at 4:59 PM, Vafa Khalighi wrote:

> Hi Ross and Jens
>
> I do not think that this is a bug of bidi and I am going to explain  
> why this is so.
>
> First Please have a look at attached tex file and its PDF output.  
> The attached PDF shows the contents of all \marginpar commands. And  
> I only had to make the left and right margin the same by using  
> \usepackage[left=5cm,right=5cm]{geometry}. So this tell me that  
> this has to do something with setting margins. Notice that when we  
> are typesetting LTR, we are writing from left to right and the  
> content of \marginpar appears on the right hand side and the  
> content of \reversemarginpar appears on the left hand side. The RTL  
> is absolutely should be opposite of this (which it is). Now in the  
> original tex file that you sent, margins are the same, I mean the  
> left and right margins are the same on , say, even pages  
> independent of RTL typesetting or LTR typesetting. So I think this  
> is natural. If you do not believe me, In the LTR part of your tex  
> file, change all \marginpar  to \reversemarginpar and see that  
> exactly the same thing that happened for \marginpar for RTL  
> typesetting in even pages, happens this time for LTR texts.

There are 2 issues here: the side on which the marginal note is placed,
as well as the directionality of its contents.

Now for positioning, using \reversemarginpar  at the
beginning of the environment indeed switches sides.
However one wonders why there should be a need for this.
With the [twoside] option, the  \marginpar  should (by default)
be placing the marginal note within the wider margin, not the
narrower one, as is currently happening.


Now marginal notes are usually used by scholars or editors,
making comments or giving references about other material.
By default LaTeX supports L to R marginal notes;
e.g., for Western scholars commenting on Western texts.

You are right in that bidi should support, by default, R to L
content in margin-pars; e.g. for Arabic/Persian scholars commenting
on Arabic & Persian texts.

However, bidi is clearly to support bi-directionality.
So it should also be making it easy for Arabic scholars to make
comments on Western texts, as well as Western scholars commenting
on Arabic/Persian texts.

That is, there should be methods within multilingual documents,
to have the directionality of margin-pars being the reverse of
that within the main body of the text.
Instead of supporting just one extra stylistic layout, there are
now a total of 2^2 = 4 possibilities, and the 3 new ones *all*
require adequate support.


To be more specific, to get the margin-pars correct in Jens'
document requires 3 things, as in the following:

  \begin{Arabic}\large
  \reversemarginpar    %   else uses wrong margin
   ... arabic content ...
  \marginpar{\LR{on p.3 \hfill}}%  for LtoR comments

   ...
  \end{Arabic}

What would be nice is to give an option to the {Arabic}
environment that would:
   a.  execute  \reversemarginpar
   b.  apply the \LR to marginpar content
   c.  remove the need for the \hfill at the end .


Even more specifically, with multi-line content --- this works:

\textbf{\small [ص ٣]}
\marginpar{\LR{on p.3 on p.3 on p.3\\ on p.3\\ on p.3 on p.3 on p.3  
on p.3 on p.3 on p.3 on p.3 \hfill}}
الخفض نحو من الله، وعلامة الفعل قد نحو  
قد قام زيد وقد يقوم والسين نحو سيقوم  
وتاء التأنيث الساكنة نحو قامت وياء  
المخاطبة مع الطلب نحو قومي، وعلامة  
الحرف أن لا يقبل شيء من ذلك.


but this does not quite do it:

\textbf{\small [ص ٣]}
\marginpar{\LR{on p.3 on p.3 on p.3\\ on p.3\\ on p.3 on p.3 on p.3  
on p.3 on p.3 on p.3 on p.3 }}
الخفض نحو من الله، وعلامة الفعل قد نحو  
قد قام زيد وقد يقوم والسين نحو سيقوم  
وتاء التأنيث الساكنة نحو قامت وياء  
المخاطبة مع الطلب نحو قومي، وعلامة  
الحرف أن لا يقبل شيء من ذلك.


nor does:  (see image)

\textbf{\small [ص ٣]}
\marginpar{\LR{on p.3 on p.3 on p.3\par on p.3\par on p.3 on p.3 on p. 
3 on p.3 on p.3 on p.3 on p.3\par}}
الخفض نحو من الله، وعلامة الفعل قد نحو  
قد قام زيد وقد يقوم والسين نحو سيقوم  
وتاء التأنيث الساكنة نحو قامت وياء  
المخاطبة مع الطلب نحو قومي، وعلامة  
الحرف أن لا يقبل شيء من ذلك.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: texshop_image.jpeg
Type: image/jpeg
Size: 83089 bytes
Desc: not available
URL: <http://tug.org/pipermail/xetex/attachments/20100201/6b559259/attachment-0001.jpeg>
-------------- next part --------------


>
> So I concluded that this is not a bug of bidi. Another solution  
> would be to use \marginpar for LTR texts but \reversemarginpar for  
> RTL texts.

In terms of defaults, no it is not a bug, and I never intended
to suggest this. Rather, there needs to be extra functionality
to properly support multi-lingual documents which contain material
set with both RtoL and LtoR directionalities.


Consider the $n^2$ growth of combinations required if vertical
typesetting (in all 4 variations of this) were to be supported also!
$n = 6$ to do them all.

>
> If you also remove twoside option from \documentclass, things will  
> be ok.

This is not really a satisfactory way to handle the issue.

>
>
> -- 
> Best wishes,
> Vafa Khalighi


All the best,

	Ross

------------------------------------------------------------------------
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
------------------------------------------------------------------------