RE: Bengali script - where is "khanda ta"?

From: Apurva Joshi (
Date: Tue May 21 2002 - 20:27:35 EDT

My responses inline. Thanks, -apurva

-----Original Message-----
From: Somnath Kundu []

Is that same for Unicode, i.e.,

Ta + Halant + Halant -> Khanda Ta

and how Uniscribe handle this case? In other words, how can I write
Ta in Unicode?

<apurva:> No, it is not the same for Unicode. The display of khanda Ta
is dependent on both: font as well as the sequence input. And if there
are additional processing layers in between-- like a shaping engine for
the script; then that needs to figure in as well. Because some shaping
engines might flag and re-order parts of sequences in order for specific
glyph substitution to be applied.
The Bangla engine in Uniscribe is based off the recommendations in
Unicode. And in case you haven't seen it, you might want to see a
specification for creating OpenType fonts for Indic scripts at: which
also includes details of shaping engine processes in Uniscribe. If you
have any feedback on it, it is very welcome.


Yes, you are right Bangla does not have distinct half forms. What I
to say is that I should be able to write Khanda Ta, Ta Halant (normally
used in Bangla words), and Ta in a conjunct in Unicode text.

Thanks for this confirmation.
Below are some input sequences, and I'm assuming that the font used for
displaying them will have a glyph for the khanda Ta. The sample conjunct
being created is taTa i.e. the double Ta. I'm using this conjunct
because it is usually diplayed as a distinct ligature.

1. To display the ligature taTa, use:
Ta Halant Ta
2. To display khandaTa Ta use:
Ta Halant ZWJ Ta
3. To display Ta Halant Ta use:
Ta Halant ZWNJ Ta


The reason for my posting was that I found Code2000 font some days ago
installed Bangla keyboard driver manually, found on my MSDN Win2k CD, on

2k/XP to type some Bangla letters but was not able to type Khanda Ta.
glyph is also probably missing in that font).

Actually, Bangla is not yet enabled on Windows. Although we do hope to
do so in a future release of Windows. Win2K has Devanagari and Tamil
enabled. WinXP has Gujarati, Gurmukhi, Kannada and Telugu enabled.

It is still not clear to me why Khanda Ta was not included in Unicode as
additional consonant, like RRA, RHA, and YYA, as it is written as a
distinct consonant (out of total 40 consonants) in Banagla script?

To the best of my knowledge, khanda Ta is not a consonant (or distinct
character) by itself, but the halant form of a consonant. Such alternate
forms can be displayed using an OpenType font, that contains glyphs for
such forms.
I don't have access to a Bangla dictionary right now, but I think the
khanda Ta is not sorted differently either; i.e. 'utkarsha' will be
listed before 'uttam'. Although utkarsha uses the khanda Ta and uttam
employs the double Ta ligature.
In case this hadn't been the case, we might then have had a case that
the khanda Ta might need more looking into.

Thanks in advance for your reply,

Somnath Kundu

This archive was generated by hypermail 2.1.2 : Tue May 21 2002 - 21:13:36 EDT