bugFree UCS Outline Fonts - Bugs: bug #41210, Introduction of new character (BA)...

 
 

bug #41210: Introduction of new character (BA) in Bangla to write Sanskrit

Submitted by:  Atul Thakur <thakur>
Submitted on:  Sun 12 Jan 2014 03:00:25 PM UTC  
 
Category: overall font problemSeverity: 1 - Wish
Item Group: glyph(s) missingStatus: Fix posted
Privacy: PublicAssigned to: Steve White <Stevan_White>
Open/Closed: OpenRelease: development

Add a New Comment (Rich MarkupRich Markup):
   

You are not logged in

Please log in, so followups can be emailed to you.

 

(Jump to the original submission Jump to the original submission)

Wed 13 Aug 2014 06:35:51 AM UTC, comment #12:

Fixes have been posted.

Again, this is a non-standard behavior included because of a perceived lacuna in the standard regarding Sanskrit.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Tue 14 Jan 2014 11:25:04 AM UTC, comment #11:

Atul: Please ignor the last post, I now see you drew it on the second page of your previous PDF.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Tue 14 Jan 2014 11:23:39 AM UTC, comment #10:

Atul, what is missing, I think, is an explanation of how conjuncts with the Sanskrit 'ba' should be rendered.

My guess is, you would have them be the same as those for the Bengli 'va', execept with the diagonal slash.

  • Is that correct?
Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Tue 14 Jan 2014 11:19:41 AM UTC, comment #9:

Vishvas: first, I should stress, I'm not sure I have time to take this on at this time. The issue remains a 'wish'.

Just the question is, whether it is appropriate to use the character U+09F0 for both purposes. Let me stress: Unicode is not about shapes of characters--it is not enough to say two characters have the same shape, to say they should be used interchangeably.

I have not found anything in the standards documentation to support this identification, but I have seen (as I listed before) several statements to the effect that the issue has not been resolved. Furthermore, the fact that the two letters behave so differently orthographically makes me doubt that the identification is a good idea.

As to how to decide this, I have to leave it to you guys of course.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Tue 14 Jan 2014 03:33:44 AM UTC, comment #8:

Steve said:

> If the right thing to do is indeed to use U+09F0 for Sanskrit 'ba', I can arrange for that to happen with reasonable difficulty (a couple of evenings of my time).


Thanks, Steve! That is very kind of you. Indeed, the exact same symbol U+09F0 is used in two different senses in bangla and assamese (in one case to represent ba distinct from va, and in another case to represent ra).

But the symbol to be used is exactly what is shown for U+09F0 .

How do we decide what the right thing to do is (use assamese ra vs map a new codepoint)? I suppose that one can do this by pinging the unicode mailing list and proposing the allocation of U+09B3 to the new "bengali sanskrit ba". (Could atul or svarUpa do this? Details are here: http://www.unicode.org/consortium/distlist.html . svarUpa has collected examples of usage here - https://sites.google.com/site/vaandbainbanglafont/usage-examples . Make sure to point to it. )

विश्वासो वासुकिजः <vvasuki>
Mon 13 Jan 2014 08:20:45 AM UTC, comment #7:

We are coming closer to an understanding, but there are still several dark areas.

Vishvas, you may want to look through the links I posted below, especially the one from Mike Mier. He seems to have thought that a separate code point was in order for the letter in question.

My opinion: Unicode is not about letter shapes, rather it is about their meanings. Only if the Assamese letter is regarded as somehow the same letter (not just the same shape) as Sanskrit 'ba' (for example, if the Assamese letter descends from older Sanskrit writings) should the same code point be used for both letters. Othewise, if the two letters just happen to look alike but have separate meanings and histories, they should have separate code points.

Atul, my remark concerning the Unicode Book v. 6.3.0:
http://www.unicode.org/versions/Unicode6.3.0/ch09.pdf
If you loook in Section 9.2 Bengali, you'll see only one note about Sanskrit--nothing regarding the issue with 'ba'. It seems to me, somebody should insert a sentence or two explaining the correct resolution of the issue.

Regarding Lohit Bengali: You have sent me version 2.5.1. My Linux distro (Ubuntu 13.10) has version 2.4.3.

These two versions behave very differently with respect to U+09F0. Version 2.4.3 treats conjuncts with U+09F0 like 'ra', producing a reph form over the paired consonant. As I recall, I thought this was correct behavior for Assamese. The current development version of FreeSans behaves like 2.4.3. Yet the new verson of Lohit Bengali treats U+09F0 as a normal consonant, and forms conjuncts rather than reph.

  • If you can, please explain the discrepancy in behavior between the two Lohit Bengali versions.

My guess is, you would like to see the behavior respecting U+09F0 as 'ra' for Assamese, and as 'ba' for Sanskrit.
This is technically feasible. I have doubts that it is the right thing to do, Unicode-wise.

If the right thing to do is indeed to use U+09F0 for Sanskrit 'ba', I can arrange for that to happen with reasonable difficulty (a couple of evenings of my time).

Otherwise, if a separate code point is in order, it would be best to await word from the Unicode consortium. Our policy has been however, that in order to include a character, we don't need a final judgment from Unicode, just an official proposal with some reasonable expectation that it may be adopted.

I mean, if we decide that a new Unicode character is the right thing, and you and your colleagues can submit a proposal to Unicode that isn't immediately rejected. I would be happy to insert the character for you.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Mon 13 Jan 2014 03:00:53 AM UTC, comment #6:

Talking with svarup and looking at http://www.unicode.org/charts/PDF/U0980.pdf , I just realized that the unicode standards do not even contain separate characters for bengali va and ba!! (They have been using the appropriate-looking assamese characters with bengali characters, and getting bad renderings of saMyuktAkShara-s.)

If this is correct, it appears that the first step is to modify unicode standards by writing to the unicode mailing list. Once there is agreement on the codepoint to be used, atul/ svarup can modify an existing opensource bengali font suitably.

विश्वासो वासुकिजः <vvasuki>
Mon 13 Jan 2014 12:39:51 AM UTC, comment #5:

Just added a pdf describing my points a bit further. I edited the image to show what behavior of BA we need which is absent in current system
Regards,
-Atul

(file #30258)

Atul Thakur <thakur>
Mon 13 Jan 2014 12:27:58 AM UTC, comment #4:

Hi Steve,
Thanks for your reply.

>>So you are not asking for a new character in the font, correct?

Yes that is correct. We don't need a new character in the font. We just need to use 0x09F0 because its shape satisfies our requirements.

>>Do I have that right?

Yes that is absolutely right. We would like 0x09F0 to behave in a particular way (i.e., in the context of Sanskrit it should form conjuncts similar in shape to 0x09AC) .

>>What am I missing here?
>>What font were you using in the image?


I think you are missing the LohitBengali font which I have added now with this post.

>>Are conjuncts in the image exactly what you expect for Sanskrit?


We expect BA (0x09F0) to behave in same way as VA (0x09AC) when it forms conjuncts with other consonants. That is to say, the conjuncts of 0x09F0 should have same shape as conjuncts of 0x09AC with a scar in the portion of glyph representing 0x09F0. The scar should be just like the scar in 0x09F0 which is absent in 0x09AC (you may look at 0x09AC having exactly same glyph as 0x09F0 without a scar).

I could not exactly follow your sentence "If your approach is really appropriate, it would be good for a sentence explaining it to appear in the Unicode Book." Is it that you are asking for a sentence of Sanskrit witten in Bangla script with appearance of VA and BA both?
Thanks for your time.
Regards.
-Atul

(file #30257)

Atul Thakur <thakur>
Sun 12 Jan 2014 07:35:31 PM UTC, comment #3:

Atul,

First, from what you say, you are using 0x09F0 for Sanskrit not because it is the officially proper character, but because it seems to have the right shape.

    • So you are not asking for a new character in the font, correct?

It seems to me this issue really ought to be made official in the Unicode standards. If your approach is really appropriate, it would be good for a sentence explaining it to appear in the Unicode Book.

Now as to what I understand you to be proposing: that the particular conjuncts that form with the 'ba' and 'va' letters should depend upon the language being used.
If that's all it is, this is quite possible, perhaps not even very hard.

    • Do I have that right?

Regarding your list of conjuncts: I made your text file into an HTML file, and opened it with Firefox (on Linux). With FreeSans, I don't see any conjuncts being formed. Yet in your image, some conjuncts are clearly being formed.

    • What am I missing here?
    • What font were you using in the image?
    • Are conjuncts in the image exactly what you expect for Sanskrit?
Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 12 Jan 2014 05:49:48 PM UTC, comment #2:

Hi Steve,
Thanks so much for your reply. I'll try to answer your questions (although I may be off due to my limited knowledge of Unicode here and for which I apologize in advance).

Firstly, I have added a file "possible_conj" that contains the consonant conjuncts with Bangla VA and corresponding conjuncts with BA (which as you rightly recognized is Assamese RA). We just mapped Assamese RA as BA due to its exact similarity with the BA character we need. Our naive approach of mapping Assamese RA to BA works well until the point we need vowel conjuncts. Henceforth, I'll refer to Assamese RA as BA for the purpose of this posting.

As you can see in the added file "possible_conj" the consonant conjuncts of BA have different shapes when compared to consonant conjuncts of VA. This is the problem. I feel this is an encoding issue rather than a font issue.

What we want for Sanskrit VA is 0x09AC.
Whereas for Sanskrit BA we need 0x09F0.
I hope that I answered the questions asked. If not do let me know.
Thanks for the links...I am going through them now.
Regards,
-Atul

Atul Thakur <thakur>
Sun 12 Jan 2014 04:23:45 PM UTC, comment #1:

Hi Atul,

It isn't clear to me what all the issues are here.

1) My understanding of Sanskrit is limited to Devanagari script. As to your list of conjuncts, are these conjuncts that must not be formed in Bengali?

    • please provide the (unicode) text file corresponding to this image

2) I don't understand the nature of the question about 'ba' vs 'va'. It isn't clear to me this is a font question, or an encoding issue.

Note: GNU FreeFont is a Unicode font: we will not add non-standard encodings. There may be other tricks we can pull, but I have to understand first.

I see that no 'va' is listed
in the Unicode 6.1 standard for Bengali
http://www.unicode.org/charts/
However, the pictures you provide show a glyph very much like
Unicode 0x09F0
BENGALI LETTER RA WITH MIDDLE DIAGONAL,
= Assamese letter ra
This character is represented already in recent versions of FreeSans.

    • Is it incorrect to use 0x09F0 for Sanskrit 'va'?
    • If not, what character is to be represented as 'va'?

Web searches reveal that this issue has been discussed many times, but I haven't seen what the resolution is.

See for instance:
================================================
Unicode 4.0 Bangla Changes - FontLab Forum
posting by Mike Meir
http://forum.fontlab.com/index.php?topic=6083.20;wap2
================================================
http://languagelog.ldc.upenn.edu/myl/ldc/IndianScriptsUnicode.html
Indian scripts and Unicode
Jeroen Hellingman
================================================
“THE SUN GOES AROUND THE EARTH”
http://arrbee.wordpress.com/2012/04/

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 12 Jan 2014 03:00:25 PM UTC, original submission:

Hello everyone,
We are trying to write Sanskrit using Bangla script. One of the problems we are facing is to express the difference between the sounds BA and VA. In Bangla, there is one one character (please see attached image VA.jpg) that represents both BA and VA. That is to say, Varun and Bakul both will be written with the character shown by attached image VA.jpg.

In Bangla this is completely okay because the pronunciation there for Varun will be Barun. However, for Sanskrit, there is a clear difference in pronunciation between BA and VA.

Can this new character BA (please see attached BA.jpg) for its glyph) be introduced in Bangla font? Also, the corresponding consonant conjuncts using VA are shown in attached file "List of the possible conjuncts.png".

A help in this matter will be highly appreciated as we are completely stuck on this point for a while and our transcription has come to a halt.

Regards,
-Atul

Atul Thakur <thakur>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

Attached Files
file #30258:  problem_conj_BA.pdf added by thakur (68kB - application/pdf)
file #30257:  Lohit-Bengali.ttf added by thakur (136kB - application/octet-stream)
file #30255:  possible_conj added by thakur (638B - application/octet-stream)
file #30252:  BA.jpg added by thakur (4kB - image/jpeg)
file #30254:  VA.jpg added by thakur (2kB - image/jpeg)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by vvasuki (Posted a comment)
  • -unavailable- added by vvasuki (shrIramaNa may be able to help svarUpa and atula with unicode standards changes.)
  • -unavailable- added by Stevan_White (Posted a comment)
  • -unavailable- added by thakur (Submitted the item)
  • -unavailable- added by thakur
  • -unavailable- added by thakur
  •  

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

    Please enter the title of George Orwell's famous dystopian book (it's a date):

     

     

    Follow 13 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Wed 13 Aug 2014 06:35:51 AM UTCStevan_WhiteStatusNeed info=>Fix posted
    Mon 13 Jan 2014 03:00:53 AM UTCvvasukiCarbon-Copy-=>Added -unavailable-
    Mon 13 Jan 2014 12:39:51 AM UTCthakurAttached File-=>Added problem_conj_BA.pdf, #30258
    Mon 13 Jan 2014 12:27:58 AM UTCthakurAttached File-=>Added Lohit-Bengali.ttf, #30257
    Sun 12 Jan 2014 05:35:20 PM UTCthakurAttached File-=>Added possible_conj, #30255
    Sun 12 Jan 2014 04:23:45 PM UTCStevan_WhiteSeverity3 - Normal=>1 - Wish
      StatusNone=>Need info
      Assigned toNone=>Stevan_White
    Sun 12 Jan 2014 03:00:25 PM UTCthakurAttached File-=>Added BA.jpg, #30252
      Attached File-=>Added List of the possible conjuncts.png, #30253
      Attached File-=>Added VA.jpg, #30254
      Carbon-Copy-=>Added -unavailable-
      Carbon-Copy-=>Added -unavailable-

    Back to the top


    Powered by Savane 3.1-cleanup