----- Original Message -----
Sent: Thursday, August 28, 2003 2:21
PM
Subject: [indic] Re: Bengali
Proposal
Dear
Sandipan,
I
spoke with the Unicode Technical Committee yesterday. We came to an agreement
that I should seek to build concensus on how to standardize Bengali before the
next UTC meeting in late September.
Being able to arrive at a concensus is critical as there is no way to
standardize implementation without people agreeing how this should be
implemented.
I am
trying to finish up the Windows update of Bengali script support. I am aware
that there is also an effort in the Linux community to have Bengali support.
Because the only contention remaining is limited to a few issues, I hope that
we can have some dialog to finally arrive at a unified approach to the Reph
and Yaphala issue.
Making the rendering of the conjuncts or ligatures the choice of the
font designer is an excellent way to address this. In OpenType there is a
"language" tag for Malayalam Repormed *and* Malayalam Traditional. I would
propose that we do something similar for Bengali, whereby a "Bengali
Schoolbook - BES" orthography can be selected that is different than
"Bengali - BEN" which would be the traditional manner in which to form
ligatures in literature. This would remove one topic from the
discussion.
In
the same manner that the previous paragraph suggests it is the font designer's
choice to support various writing styles, it should also therefore be the
choice of the font designer to either ligate or not ligate the yaphala without
the use of ZWJ. As above, the OT language differentiation could also allow
different behaviors in the same font. This would remove another issue from the
discussion
That
then leaves only the Reph and Yaphala item and Khanda Ta item as issues that
need to be resolved.
REPH
and YAPHALA
Since the YA with REPH above is more commonly used, the RA HALANT YA
should produce that result. I have proposed that the RA YAPHALA be formed by
using RA ZWNJ HALANT YA.
By
Unicode rules, Cons-Halant forms the half form. The REPH is the half form of
the RA at the beginning of a cluster.
By
Unicode rules, Cons-Halant-ZWJ forces the formation of the half form. This
would indicate that the REPH should be formed at the beginning of a cluster as
well.
By
Unicode rules, Cons-Halant-ZWNJ forces the halant form. This would have the RA
HALANT displayed.
The
logic of using the RA ZWNJ HALANT was to keep the RA HALANT apart so that the
YAPHALA COULD be formed as the postbase of the syllable.
Is
it possible to gain a concensus on the Unicode character stream behavior for
the REPH and YAPHALA based on the logic presented above?
Regards,
Paul
Nelson
Hello Paul,
We came across your proposal on Bengali
Unicode. First, we appreciate your efforts to standadize the implementation
of Bengali script on Unicode. Working on this for the past two
years, we have a first hand experience how lack of standardization
is weakening the collective efforts.
Based on our experience of working with
the Bengali Unicode script, I have responded to your proposal and your
concern on the changes suggested by Bangla Schoolbook Reform of Bangladesh in
the attached PDF document. Please let me know whether this helps you. We would
be eager to participate in your process of formulating the standard. Lets keep
the communication channel open.
Thanks and regards.
--sandIpan
Sandipan Sarkar
Celcius Technolgies
----- Original Message -----
Sent: Thursday, July 03, 2003 12:19 PM
Subject: [Indic] Bengali Proposal
> Dear Sir,
>
> Please see the forwarded
mail from MR. Paul Nelson, regarding the Unicode
> for Bengali Script.
Kindly send your comments to Unicode discussion forum
> with a copy to
Department of IT, New Delhi.
>
> With regards,
>
>
> Manoj Jain
> Scientist "C"
> Government of India
>
Ministry of Communications & IT
> Department of Information
Technology
> New Delhi 110003
> Phone +91-11-24301240 Fax
+91-11-24363076
>
>
> ----- Forwarded by Manoj Jain/doe on
07/01/03 01:32 PM -----
>
>
"Paul
Nelson
>
\(TYPOGRAPHY\)"
To: <address@hidden>, "Omi
Azad"
>
<address@hidden <address@hidden>, "Solaiman
Karim"
>
rosoft.com>
<address@hidden>, "Peri
Bhaskararao"
>
Sent
by:
<address@hidden>, "Sanjiv
Burman"
>
address@hidden <address@hidden>, "santosh
rastapur"
>
de.org
<address@hidden>,
"Shamsuddoha/T&ITBangladesh"
>
<address@hidden>, "Mike
Meir"
>
<address@hidden>, "Naushad
Jamil"
>
07/01/03 10:41 AM
<address@hidden>,
"Andy"
>
<address@hidden>, <address@hidden>
>
cc: "Cathy Wissink" <address@hidden>,
>
"Raveesh Gupta" <address@hidden>
>
Subject: [indic] Bengali
Proposal
>
>
>
>
>
> The attached is an update to the
document I wrote, L2/03-209 posted in the
> Unicode document register. I
am including this to a wider distribution than
> the Unicode list in
hopes of being able to get some type of
> resolution/concensus on how
Unicode should work with the Bengali script.
>
> Following are
some observations and presuppositions that go into my
>
document.
>
> 1. This document is not about input sequences or
methods. It is about
> defining the rendering behavior of Unicode
character streams when they are
> used in documents.
>
> 2.
It is understood that typographic preferences exist between various
>
parties. This document is to make sure that those preferences are
>
addressable in plain Unicode text (i.e. text without any higher level
>
markup).
>
> 3. It is understood that some examples shown do no
exist in real life.
> However, as an implementer of an engine for
Bengali script I need to
> understand how to respond and advise font
makers to code their fonts if an
> "non-existent" use of codepoints is
encountered.
>
> It is my sincere desire that this document will
allow a unified use of
> Unicode encoded text that will stimulate the
global transfer of Bengali
> script text on any platform.
>
> All behavior not mentioned in the document should be assumed to
follow
> documents previously published by the Unicode Consortium. I
have taken
> special care to follow the Unicde rules and
definitions.
>
> Please forwrd the attached document to others
who should also be included
> in this discussion.
>
>
Regards,
>
> Paul Nelson
> Development Lead
>
Microsoft Font Technology(See attached file:
>
BengaliProposal_2003_06_30.pdf)
>
---
Outgoing mail is certified Virus Free.
Checked
by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.512
/ Virus Database: 309 - Release Date:
08/19/2003