W3C home > Mailing lists > Public > xsl-editors@w3.org > January to March 2003

Clarrification Needed: Implications of writing mode on region-body

From: W. Eliot Kimber <eliot@isogen.com>
Date: Wed, 12 Mar 2003 08:48:00 -0600
Message-ID: <3E6F48A0.408@isogen.com>
To: xsl-editors@w3.org
CC: support@antennahouse.com, xep-support@renderx.com


There is disagreement between XSL Formatter and XEP over the implication 
of setting writing-mode on region-body and my reading of the spec cannot 
find a clear justification for either interpretation.

In section 7.27.7, the definition of the writing-mode property, the spec 

o When "writing-mode" is applied to the fo:region-*, it defines the 
column-progression within each region. The inline-progression-direction 
is used to determine the stacking direction for columns (and the default 
flow order of text from column-to-column).

o To change the "writing-mode" within an fo:flow or fo:static-content, 
either the fo:block-container or the fo:inline-container, as 
appropriate, should be used.

With XEP, setting writing-mode on region-body only changes the positions 
of the columns on the page--it does not affect the writing mode of the 
areas within the flow contained within the region body.

With XSL Formatter, the writing mode is propogated to the areas within 
the region body.

XEP's approach is supported by two pieces of evidence I can find:

- The fact that different page masters can have different writing modes 
for region-body, which would only make sense if the intent is to only 
affect the position of columns (enabling inside/outside column position 
layouts, which are otherwise impossible to achieve with XSL-FO 1.0).

- The two bullets above, which can be read as limiting the affect of 
writing mode on simple-page-master and region-body, effectively creating 
a special case that overides the normal propogation rules for the 
writing-mode property (that is, that the writing mode of a containing 
reference area (region-body) determines the writing mode of normal areas 
within that reference area (areas within the flow)).

XSL Formatter's approach is supported by the fact that writing-mode is 
set on reference areas and that there is no other direct way to set the 
writing mode on a page sequence or flow except to set it on the 
body-region (which is the nearest ancestor reference area for the flow 
areas). The only other way to do it would be to put the entire flow into 
a block-container that sets the writing mode [see note below]. Also, the 
two bullets above do not explicitly say that they are defining the 
*only* effects of writing-mode when specified on simple-page-master or 
region-body, so they can be read as defining *additional* implications 
in addition to the normal implication of writing-mode on reference areas.

Thus, my question is this:

Does setting writing-mode on simple-page-master or region-* affect 
*only* the placement of page regions or columns or does it also 
determine the default writing mode of any contained areas?

If I hadn't thought about the implications of writing mode being 
different for different page masters or body regions within a page 
sequence, I would have assumed that writing mode on page masters or page 
regions sets the writing mode for their contained areas (XSL Formatter's 
interpretation), but the fact that one could have different writing 
modes on different page masters calls this into question, since the only 
reason for having different writing modes on different pages would be to 
control area and column position but not text writing mode (since one 
would not normally want writing mode to change based on the chance break 
of a page and if you really needed to control writing mode that closely 
you could use block containers within the flow).

Note that requiring a root-level block-container to set writing mode 
would seriously complicate the markup for multi-column layouts that 
require column spanning (which is all of them in my experience), as you 
would have to intersperse flow-child block containers with flow-child 
blocks with span="all". This would make generation of flow content quite 
difficult at best.

This analysis suggests either that XSL Formatter's interpretation is 
correct (root-level block-container not required to set flow-level 
writing mode) or that the spec is underspecified and should provide a 
way to set writing mode for the flow's normal areas at the flow or page 
sequence level in a way that does not interfere with writing-mode set on 
page-sequence-master or region-body or that having different writing 
modes on different region-bodies within the same page sequence should be 
a reportable error (and thereby avoiding the problem altogether).

This issue was raised by an XEP user's attempt to get an inside/outside 
page layout that puts headers and marginalia in the outside column and 
content in the inside column. This could be done *if* XEP's 
interpretation is correct, otherwise it cannot be achieved in XSL 1.0. 
However, because columns must be of equal width, this solution would be 
unsatisfactory for most use cases (where the content column would 
normally be wider than the header/marginalia column). Therefore, I do 
not find this use case to be a compelling argument for XEP's 
interpretation. I also note that XEP provides extensions to fo:float 
that allow inside/outside float positions, which does enable a 
satisfactory solution to this particular layout requirement (it could 
also be satisfied by providing inside/outside positioning of list item 
label and body).


W. Eliot Kimber, eliot@isogen.com
Consultant, ISOGEN International

1016 La Posada Dr., Suite 240
Austin, TX  78752 Phone: 512.656.4139
Received on Wednesday, 12 March 2003 09:48:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:44:22 UTC