Material for discussions on issue OFFICE 3854

[Regina Henschel <regina.henschel@libreoffice.org>]

Hi all,

this is for https://issues.oasis-open.org/browse/OFFICE-3854.

Inspired by the warning of Camilla Boemann to "be absolute certain the 
meaning is exactly as we want it", I have investigated how LibreOffice 
and Apache OpenOffice currently calculate percentage values.

I have found two used practices:
(1)
This way corresponds to the CSS box model as described in 
http://www.w3.org/TR/CSS21/box.html. What there is called the "content 
area" is used when the picture is

* anchored to page
The outermost box is the page for paged media, and the window into which 
the document is rendered for continuous media.

* anchored to frame
The outermost box is the frame itself.

(2)
The reference area described here is used as base for the percentage in 
all cases not in (1). That are all cases where a picture is anchored to 
paragraph, to character or as character.

Width:
The width of the reference area is the text column width. I mean those 
columns which you can set for pages or sections. Such columns can be 
defined too in a table cell or merged table cells and for headers or 
footers by using a section. The gap between the text columns do not 
belong to the area. This description includes the case of having only 
one text column.

Height:
case anchor is in section:
	Take height of that part of the section, which is on the same page as 
the anchor.

case anchor is not in a section:
	case page:    Take height of content area, subtract height of header 
and footer.
	case frame:   Take height of content area.
	case table cell: Take row height, subtract border part of the cell, 
subtract padding.

continue for all cases:
Subtract the height of the footnote area of that column in which the 
anchor lays. The footnote area height consists of the thickness of the 
footnote separator, the space above and below the separator and the area 
height of the footnotes themselves.

The UI of LibreOffice and OpenOffice gives "text boundaries" for this 
reference area. Such text boundaries are shown in Magenta in some of my 
attached screenshots.

I see three tasks:
(A)  Find suitable terms for the values of the new attribute.
(B)  Improve the description in 19.509 and 19.510.2.
(C)  Create a glossary of technical terms. Such glossary would be useful 
for readers and would help the editors to be consistent throughout the 
specification text. Whether to integrate such glossary into the 
normative part of the specification is a different topic.

Kind regards
Regina

Attachment: Issue3854_current_in_LO_AOO.odg
Description: application/vnd.oasis.opendocument.graphics