Weak characters handling in OO

[Waleed Hosny <waleed dot hosny at sun dot com>]

Hi All,

I would like to introduce myself, I am Waleed working for Sun 
Microsystems and currently there are undergoing a discussion about 
handling weak characters in OO documents.
I would like to have your feedback on this issue and what is the best 
way to handle it based on your valuable experience with this matter that 
you have developed.

Sincere Regards,
Waleed

----------------------------------------------------------------------------------------------------------------------------
Hi All,

in the successor of OO.org 2.0 we must address the issue of the 
direction of weak characters.
This mail refers to IssueZilla issue 18024.

for the moment we have to options to deal with the direction of weak 
characters in OO.org's next release (after 2.0)


1. Behave like MS Office does and determine the direction by reading out 
the locale of IME used to input the weak character. So any weak 
character input with a R2L locale will be R2L and vice versa.

Pros:
o 100% compatibility with MS Office documents that use BiDi text.
o Same behaviour like MS Office on BiDi text when inputting

Cons:
o Currently only Windows IME report their locale, Unix' IMEs don't.
 Result: OO.org for Windows will behave differently to OO.org for
         Unix
o Apparently the MS Office method is not correct under all circum-
 stances, so we will copy MS bugs into OO.org



2. Behave like suggested in issues 18024 by introducing a modification 
to the Unicode standard.

Pros:
o Better and more correct input on weak characters, beating MS Office
o Being independent from IME capabilities since no read-out is necessary

Cons:
o Not compaible to MS Office file format
o Not compatible to MS Office input


Given that MS Office has a market share of well above 90% in all markets 
in question the MS Office interoperability< becomes a very serious issue 
we must address.
What we (Star Office) can think of for the moment is to implement an 
additional compatibility flag that treat MS Office documents correct on 
import and export.
When opening a OpenOffice.org document we will use the enhanced version 
of Unicode input.
But that would mean that, whenever editing a MS Office document, the 
user will experience a different way of input compared to a native 
document.
Furthermore under Unix the input also might be completely off since we 
have no locale support under these OS IMEs.

so basically what we must discuss here is a viable solution for all 
stakeholder - and that's you, guys :o)

Please state your suggestion and best compromises you can think of.

Thank you very much in advance.

Regards
Falko Tesch