Re: [office] Bi-directional text.

[Lars Oppermann <Lars.Oppermann@Sun.COM>]

Daniel Carrera wrote:
> Lars Oppermann wrote:
>> You can have LTR text in a RTL paragraph by using left-to-right-mark 
>> and right-to-left-mark or RTL/LTR-embedding unicode characters in the 
>> text.
> 
> What part of the spec specifies this? I just did a search and there 
> aren't any instances of the strings "right-to-left" or "left-to-right" 
> on the spec.

This is not part of the OpenDocuemnt specification. It is part of the 
unicode standard which is referenced by OpenDocument (The whole unicode 
standard, not just this part of course).
See http://www.unicode.org/versions/Unicode4.0.0/ch15.pdf under 15.2: 
Layout Controls, Bidirectional Ordering Controls. These can be used to 
control the direction in mixed (western/arabic) content. For this to 
work correctly, the base direction of the enclosing paragraph needs to 
be established, which is supported by OpenDocument Section 15.5.36, 
style:writing-mode attribute which specifies the default writing 
direction of a paragraph.
See also:
Section 15.2.19 for page base direction (e.g. column layout)
Section 15.7.8 for sections
Section 15.8.13 for tables

This base direction can then be used in conjunction with the unicode 
bidi-algorithm (http://www.unicode.org/unicode/reports/tr9/) and the 
marks/signs from unicode-15.2 in order to display the content correctly.

> Jonathan's email says that character encoding is not enough (I don't 
> know if what you described is character encoding). I can't really say if 
> the present situation is enough, because I can't find the portions of 
> the spec that you are referring to.

True, character coding alone isn't enough, but the combination of the 
attributes that are available in OpenDocument (see above) to establish a 
base direction in combination with the mechanisms provided by the 
unicode standard cover the requirements of which I am aware.

> I notice that when ODF was approved at ISO a lot of comments were about 
> insufficient Bidi support. Do you think this is because the ISO members 
> didn't realize which parts of the spec gave proper bidi support?

I think that the problem might be due to the fact, that a) there is no 
explicit section in OpenDocument describing all bidi functions and b) 
the functionality is a combinations of attributes defined in 
OpenDocument and mechanisms specified in unicode. It's hard to keep 
track of things that way. Maybe a set of examples and their appropriate 
coding in OpenDocument would be a useful resource for implementors as 
well as evaluators of the OpenDocument standard. Any volunteers?

All the Best,
/lars