bugFree UCS Outline Fonts - Bugs: bug #54997, Bengali incorrect ligatures


bug #54997: Bengali incorrect ligatures

Submitter:  Mahfuzur Rahman <burhandanger>
Submitted:  Fri 09 Nov 2018 07:42:03 PM UTC
Category:  individual character(s) Severity:  3 - Normal
Item Group:  character substitution issue Status:  Proceeding
Privacy:  Public Assigned to:  Stevan_White
Open/Closed:  Open Release:  * development
* Mandatory Fields

Add a New Comment Rich Markup

Jump to the original submission

Thu 29 Nov 2018 07:42:22 AM UTC, comment #9: 

In todays commit (serif and sans normal only), I implemented two font features 'dlig' and 'hlig', for a "transparent" style and a "traditional" style, respectively.  The former turns many special ligatures off, and replaces a few others; the latter turns on ligatures with -ya.

By default, the set of ligatures should now be nearer to my judgment of a reasonable common set of ligatures that are produced in modern Bengali fonts.

I have not yet carefully checked that this addresses all the issues you raised, but I have run the fonts through some of the tests I have previously used.

Steve White <Stevan_White>
Group administrator
Sun 18 Nov 2018 10:48:33 PM UTC, comment #8: 

There have been a few quite incorrect ligatures, but mostly, "incorrect" is an oversimplification of the situation.  But I had misunderstood the situation.

Many of the ligatures you're complaining about have been common in printing in the recent past.  And there is no official set of ligatures in common use.  It also seems to be clear that different people have different ideas of what a proper set of ligatures is -- from an imagined traditional set of ligatures, to various ideals of a simplified system.

And then there is a middle road taken by most modern fonts, which take many of the "traditional" ligatures, but exclude certain others.  But I've looked at eight or ten fonts, free and commercial, and find substantial differences among the ligatures they support.

For this project, I prefer a conservative approach, for unsurprising reading of documentary text -- but a conservative approach should not force glyphs that appear antiquated.

On the other hand, we want to support older documents as well.  And a simplified script does make sense in some applications.  Fortunately, OpenType has features that allow the choice of a style of these glyphs.

In today's commit, I've disabled the traditional -ya ligatures that you complained about, unless the font's 'hlig' feature is activated.  And if the 'dlig' feature is activated, many combinations, such as the special -ukaar and -uukaar vowel combinations (among others), are disabled.

So far, I have worked on the normal faces only.

Steve White <Stevan_White>
Group administrator
Sat 10 Nov 2018 11:15:06 PM UTC, comment #7: 

You will find the conjuncts with ya on page 40 of this document:

amar bornomala o somosamoyik bangla typography

But the document is in Bangla.  Perhaps you could summarize for us what the author is saying about the ligatures.

Steve White <Stevan_White>
Group administrator
Sat 10 Nov 2018 08:16:45 PM UTC, comment #6: 

About point (3), I'd say if there's a fix for (1) there should be one for (3) or there shouldn't be a fix for (1).

1 and 3 are basically same thing.

'da' with 'short u' changes  'da' and 'da' with 'long u' doesn't.
'sa' with 'short u' changes  'sa' and 'sa' with 'long u' doesn't.

This is just creating inconsistencies.

about point (5), I'm sorry. I couldn't make my points clear. I meant these doesn't look like sans to ME. Sans is meant to be clearer if I'm not wrong. I dislike the way how ' ্য' connect with preceding letter at the bottom. I'm not going to argue much since I'm the only one complaining.


হ্য দ্য and other characters might bother me, but I'll learn to live with it.

Thanks !

Mahfuzur Rahman <burhandanger>
Sat 10 Nov 2018 06:53:07 PM UTC, comment #5: 

I just committed changes meant to resolve your issues 1, 2, 6 and 7.

I think all the other issues regard forms that I understood to be "traditional".

Have a look!

Steve White <Stevan_White>
Group administrator
Sat 10 Nov 2018 02:45:35 PM UTC, comment #4: 

As you say, issues 6 and 7 are outright bugs.  I see where the problem is, and will fix it.

Regarding issue 1: I agree it seems unlikely the shape of the letter da should change with the ukaar.  I will likely just fix this.  But I want to survey the resources at hand.

Regarding 2 and 3: this may be a mistake or misunderstanding on my part -- the vowel appears to be attached to the "half" form of sa. Again I'll survey my resources.

Regarding 4: It is my undestanding that these ligatures are traditional forms, common not too long ago in Bengla typography (which is probably what you mean by "standard"). I have heard of a reform, associated with the word "transparent", that removed many such conjuncts.  On top of this, of course, I may have misunderstood things or made technical mistakes.

I will need to understand the current state of affairs better, and the history, before taking action here.

I'm not sure what you mean by issue 5 -- I guess it's that the consonant has changed form somewhat in these letters in Sans, but not in Serif.

Thanks again!

Steve White <Stevan_White>
Group administrator
Sat 10 Nov 2018 09:24:14 AM UTC, comment #3: 

> Wrong

I've said দ্য to be wrong. I'm confident that if this's not placed alongside Bangla characters bengali people won't be able to recognise this.

> Not Standard

I agree I misused this. I've taken into account that 12 grades of academic books that come from government book and national newspaper. I've taken them as a standard. Actual standard is probably set by 'Bangla Academy' and I've not gotten into that.

Mahfuzur Rahman <burhandanger>
Sat 10 Nov 2018 09:03:31 AM UTC, comment #2: 

> Could you please post a PDF version? That should capture what you are seeing.

I've attached both .pdf and .odt. You may have missed that.

It's https://savannah.gnu.org/bugs/download.php?file_id=45394

Mahfuzur Rahman <burhandanger>
Sat 10 Nov 2018 08:52:08 AM UTC, comment #1: 

Hi Mahfuzur,

Thanks for this.

We first need to be sure that we are seeing the same thing.
Word processor files are especially bad for this purpose (they are intended for desktop publishing, not for exact transfer of documentation).

Could you please post a PDF version?  That should capture what you are seeing.

I have obtained a copy of SolaimanLipi... but so far, I am perplexed.  In some cases, it produces shapes to my eyes very much like FreeSerif, on my system anyway.  I wonder if all the issues are font problems, or something else.

We need to be careful with our terminology.  Here are some suggestions:

-- "wrong" should mean nobody ever wrote or printed it that way, and it is unreadable.
-- "old fashioned".  We can work with that technically.
-- "not like my favorite font".  This is of little interest.
-- "not standard" to mean there is an official standard to be referred to.
-- "ugly" should mean readable, but odd to the point of disturbing reading

You also used the description "stylist font".  A fancy special letter of combination is also something we can deal with technically.

Steve White <Stevan_White>
Group administrator
Fri 09 Nov 2018 07:42:03 PM UTC, original submission:  

Please Check the attachment.

Mahfuzur Rahman <burhandanger>


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

Attach Files:

Attached Files
file #45403:  page 40.png added by burhandanger (309KiB - image/png)
file #45394:  bangla.pdf added by burhandanger (46KiB - application/pdf)
file #45395:  bangla.odt added by burhandanger (19KiB - application/vnd.oasis.opendocument.text)


Depends on the following items: None found

Items that depend on this one: None found


Carbon-Copy List
  • -email is unavailable- added by Stevan_White (Posted a comment)
  • -email is unavailable- added by burhandanger (Submitted the item)

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only logged-in users can vote.


    Follow 7 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2018-11-18 Stevan_White StatusNeed info Proceeding
    2018-11-11 burhandanger Attached File- Added page 40.png, #45403
    2018-11-10 Stevan_White Item Groupglyph(s) incorrect character substitution issue
        StatusNone Need info
        Assigned toNone Stevan_White
    2018-11-09 burhandanger Attached File- Added bangla.pdf, #45394
        Attached File- Added bangla.odt, #45395

    Back to the top

    Powered by Savane 3.13-b921.
    Corresponding source code