RE: Sanskrit nasalized L

From: Peter Constable <>
Date: Sat, 18 Jun 2011 17:26:24 +0000

When support for Devanagari was first added to Windows, the input coming from Indian experts that drove the design had the perspective of the major modern languages only. From that perspective, candrabindu and visarga are in complementary distribution, and so the design allowed only one to occur. From my way of thinking, the mindset for implementation of this and other scripts was inappropriately focused on orthographic conventions for particular languages and not on script behaviours. In the Vista and Windows 7 releases, we took significant steps to relax constraints in our shaping behaviour for various scripts, but we later learned about some additional issues. So, while we don't yet support all of the Vedic marks added in Unicode 5.2, we are making progress in supporting most other sequences that may be found in Vedic text as well as less common sequences found in modern usage.


-----Original Message-----
From: [] On Behalf Of Shriramana Sharma
Sent: Friday, June 17, 2011 6:41 PM
Subject: Re: Sanskrit nasalized L

On 18-06-2011 00:40, Vinodh Rajan wrote:
> BTW I am still not able to display आँः& वाँः without the dotted
> circle in Windows 7:-)

Ouch! I'ven't had the chance to try out the rendering on Win 7. But it seems it's still problematic.

Last time I complained about this on Unicore a couple months back Peter Constable said "try out our latest software and then get back" -- maybe he meant the developer version of Windows 8?

So when this kind of fudged rendering happens, font developers are forced to create hacks in their rendering tables or other software -- Sanskrit 2003 uses a separate (non-combining) visarga in the PUA. Elmar Kniprath's fonts use all sort of (strictly speaking) unnecessary substitutions in the OT tables.

I would be willing to work with Microsoft to help them achieve proper Vedic rendering correctly. I am already on the Harfbuzz mailing list for FOSS, but then Windows is a different beast as it is closed-source.

Shriramana Sharma
Received on Sat Jun 18 2011 - 12:28:55 CDT

This archive was generated by hypermail 2.2.0 : Sat Jun 18 2011 - 12:28:56 CDT