DBCS/ SBCS font issue   

The Vrinda font (a Microsoft-supplied Unicode font for the Bengali code 
range) is installed on both my office and home PCs.  (The Ms Arial 
Unicode font also has characters in the Bengali range of Unicode, but 
not all of them render correctly; the Vrinda font, which I believe came 
with SP2, fixes the problem.)

Both PCs are running Windows XP SP2 and Office 2003, and both have 
complex script support turned on in the 'Regional and Language Options' 
dialog box in the Control Panel.  And both will use the Vrinda font for 
non-office applications, e.g. Notepad.  But under MsWord, only my office 
PC displays Bengali characters; Word in my home PC displays them as 
boxes.  In fact, my home Word acts in some ways as if the Vrinda font is 
not installed: if I select Bengali text, Word tells me it's using the 
'DevanagariMT' font.  This is bizarre, because Bengali does not in fact 
use the Devanagari script.

I've looked for some setting I might have wrong.  The closest I came is 
in the Tools | Options dialog, Compatibility tag.  If I click on the 
'Font Substitution' button, I get yet another dialog box that tells me I 
am missing the 'DevanagariMT' font, and that it will substitute the 'MS 
Mincho' font.  OK, I thought, I'll play along: I told it to instead 
substitute the Vrinda font for this so-called 'DevanagariMT' font.  No 
luck, but maybe a clue: it tells me
    You can't replace DBCS font with SBCS font or
    SBCS font with DBCS font.
My understanding is that SBCS would be an ISO encoding.  This is all 
supposed to be Unicode, so I don't know where Word is getting the idea 
that I have SBCS text in there.  When I paste it in to other programs 
(like Notepad, or jEdit), they confirm what I paste is Unicode text (and 
at least jEdit confirms that it is UTF-8).

On the other hand, I can insert a Bengali character in the Vrinda font 
into this document using the Insert | Symbol command.  And closing and 
reopening the file (as a Word doc) preserves such characters.  So in 
some ways, it looks like an error in the way my office PC's document is 
transferred.  In this particular case, I emailed myself the Word 
document, sending it from Outlook and receiving it from Thunderbird. 
But that hasn't been a problem in the past, because I've sent plenty of 
other Unicode documents back and forth that way to other applications. 
And I've also seen the same issue with documents that are exported from 
another program (XMLmind) in WordXML format: Bengali characters are fine 
when I open the exported document on my office PC, but show up as boxes 
on my home PC, despite the fact that I exported the documents on my home 
PC, so they never went through the mail.

Any ideas why Word behaves differently on the two systems?

    Mike Maxwell

date: Tue, 18 Sep 2007 20:52:51 GMT   author:   Mike Maxwell