xsl-before-float-separator & xsl-footnote-separator

The editors,

The behaviour of the above classes of static-content in the draft 1.1 
Recommendation of 16 December 2004 has not changed, as far as I could 
see, from the 1.0 Recommendation.

There are some problems that are peculiar to these flows.  In XSL 1.0, 
the combination of implicit flow-map, uniqueness of region-name within a 
simple-page-master and the uniqueness of flow-names guarantees, if I am 
not mistaken, that any given region on a given page will be the target 
of at most one flow, be it fo:flow or fo:static-content.  In combination 
with the method of determining the dimensions of the region-body and the 
extents of the "border" regions, this ensures that the layout available 
to the fo:flow does not have to be adjusted depending on the contents of 
any fo:static-region.

If retrieve-markers occurs in xsl-footnote-separator or 
xsl-before-float-separator, however, they directly impact the space 
available for the layout of fo:flow.

This situation is different from that pertaining to 
fo:retrieve-table-marker, which more closely parallels the treatment of 
footnotes than that of marker retrieval in that the fo context is 
contained entirely within the fo:flow.

The draft and the Recommendation provide for ad hoc restrictions on a) 
whether a conditional area is generated at all if there is insufficient 
space in the main-reference-area, and b) on the total b-p-d space that a 
conditional area can occupy.  These clauses do not seem to apply to the 
separators in isolation.

If a survey of implementations revealed that marker retrieval was not 
implemented within xsl-before-float-separator & xsl-footnote-separator 
static-content, would there be any interest within the working group in 
constraining such static-content to exclude fo:retrieve-marker?

Yours faithfully,
-- 
Peter B. West <http://cv.pbw.id.au/>
Project Folio <http://defoe.sourceforge.net/folio/>