bugFree UCS Outline Fonts - Bugs: bug #45157, Pe̍h-ōe-jī: O with right dot...

 
 

bug #45157: Pe̍h-ōe-jī: O with right dot and other diacritics positioned incorect

Submitted by:  Hsing-yen He <hammerandsickle>
Submitted on:  Thu 21 May 2015 05:21:18 AM UTC  
 
Category: individual character(s)Severity: 3 - Normal
Item Group: letter positioning poorStatus: Fix posted
Privacy: PublicAssigned to: Steve White <Stevan_White>
Open/Closed: OpenRelease: development

Add a New Comment(Rich 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 30 Sep 2015 07:13:42 AM UTC, comment #36:

This diacritic should works with dotted circle.

Anonymous
Sat 29 Aug 2015 09:16:54 PM UTC, comment #35:

OK, in the 20150829 build these appears better than past builds. Thanks very much.

Hsing-yen He <hammerandsickle>
Sat 29 Aug 2015 12:49:34 PM UTC, comment #34:

Hi,

I got the test file from the last comment, they seems also work with me, but still having a small problem when I turned small caps on:

In all faces of FreeSans, the right dot looks higher than lowercase.
In FreeSerif Italic and Bold Italic, the right dot seems have a bit problem.
FreeMono seems no problem for me.

Hsing-yen He <hammerandsickle>
Sat 29 Aug 2015 10:41:08 AM UTC, comment #33:

Please try it with the improved file
A_test_for_Pe̍h-ōe-jī_letter_rending.htm
The positioning seems to work for me.

(file #34763)

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 15 Aug 2015 07:45:16 PM UTC, comment #32:

There is still having problem on when I turned small capital on, tested in this page.

Hsing-yen He <hammerandsickle>
Wed 22 Jul 2015 03:31:13 PM UTC, comment #31:

OK, the combining dot above right was placed corretly in all faces of FreeFont, this bug is fixed.

(BTW if you fixed i̍ in some faces of FreeFont, the font family will have advantages to compete with other font families that providing support for Pe̍h-ōe-jī.)

Hsing-yen He <hammerandsickle>
Tue 14 Jul 2015 10:59:37 PM UTC, comment #30:

I have already noticed that other typefaces of FreeFont was already implemented it. At first, I compared any faces of FreeSerif, found the height of right dot in FreeSerif Bold, FreeSerif Bold Italic and FreeSerif Italic looks higher than FreeSerif Regular. Then I found in FreeSans the right dot also higher, FreeMono only uppercase letters get proper typesetting. I.E. Only FreeSerif Regular got proper typesrtting for dot above right, other typefaces need one step to get fully resolved.

Hsing-yen He <hammerandsickle>
Tue 14 Jul 2015 12:29:16 AM UTC, comment #29:

Sure, it's other faces of FreeFont, I found the height of dot above right looks different from FreeSerif Regular when I compare with them carefully. I think this issue is only one way to be resolved.

Hsing-yen He <hammerandsickle>
Mon 13 Jul 2015 09:34:44 PM UTC, comment #28:

Hsing-yen,

Which "other faces"?

The other faces of FreeFont? But it is already implemented there.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Mon 13 Jul 2015 05:38:33 PM UTC, comment #27:

Hi Steve,

OK, I means the solution in FreeSerif Regular is really suitable for other faces. I think this issue is almost resolved.

Hsing-yen He <hammerandsickle>
Mon 13 Jul 2015 11:48:26 AM UTC, comment #26:

Hi Hsing-yen,

Do you mean that now you prefer the solution in FreeSerif to the solution in other faces? (I don't really understand your sentence in the last post.)

Can we say this issue is resolved?

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Mon 13 Jul 2015 11:18:14 AM UTC, comment #25:

Well, in my opinion, only FreeSerif Regular performanced desirable now; in other typefaces, this issue can be resolved just through adjusting the height of dot above right as FreeSerif Regular behaviored.

Hsing-yen He <hammerandsickle>
Mon 29 Jun 2015 07:32:52 AM UTC, comment #24:

Steve,

Now I think the reason that this confilict appeared is maybe their publishers encounterd technical limit. Despite this diacrical was introduced in Unicode for a long time, but O with right dot is "not well-supported by fonts", so publishers tend to using a spacing modifier letter instead.

I also think you have set an good example in FreeSerif Regular that other typefaces can catch up with.

Hsing-yen He <hammerandsickle>
Thu 04 Jun 2015 09:38:38 PM UTC, comment #23:

Hsing-yen,

Please respond to my comment #17.

No action will be taken until we understand why the Unicode standard and your images seem to be in confilict.

It is geometrically/typographically impossible to do what you are asking. If I push the dot any lower, it will collide with the o at normal resolution. It surprises me that this isn't obvious to you -- which makes me wonder if we're seeing the same thing.

As you remember, we ignore comments about how other fonts behave.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Mon 01 Jun 2015 03:25:03 AM UTC, comment #22:

Hi Steve,

Thanks for your advise. In fact, the spacing style is not essential in these sequences, you can just fixing the height of right dot (except FreeSerif Regular). I have also tested with Linux Libertine, this font family have no spacing style when I added combining dot above right, but they looks porformanced better then FreeFont in some sequences.

Hsing-yen He <hammerandsickle>
Sun 31 May 2015 06:39:18 PM UTC, comment #21:

Hello, Hsing-yen,

It is good that you acknowledge that there is a discrepancy between the images and the Unicode description of this character.

No, I won't make arbitrary changes, only to find later that the mark really needs to be a spacing mark.

As I wrote in comment 17, before we proceed, we must understand how it is that Unicode specifies this character as a non-spacing mark above the letter, while your images show a spacing mark beside the letter.

Please re-read comment 17 carefully. I thought about it a lot.

I do not have the resources to discover this -- I think you do, however.

Once the situation is explained, we can consider what to do about it.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 31 May 2015 06:04:25 PM UTC, comment #20:

Oh, sorry, I have reviewed some attached files, and you're right that they're showing spacing mark. But I still insist that the height of right dot should be reduced like FreeSerif Regular.

Hsing-yen He <hammerandsickle>
Sun 31 May 2015 07:20:31 AM UTC, comment #19:

Hi Hsing-yen,

Are we talking about the same thing? I'm talking about the dot that follows the o like a separate letter in these examples:

(file #34134)

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 31 May 2015 03:12:18 AM UTC, comment #18:

Well, in my opinion, the "lower" means to reduce the hight of right dot, as FreeSerif Regular shown.

Hsing-yen He <hammerandsickle>
Sat 30 May 2015 06:07:21 PM UTC, comment #17:

Here is the entry from the file
http://www.unicode.org/Public/7.0.0/ucd/extracted/DerivedCombiningClass.txt
reads

0358 ; 232 # Mn COMBINING DOT ABOVE RIGHT

(This is the Unicode 3 version. The 7.0 versions say the same thing, but are not so nice to paste into a bug report.)

The third field is a character type description. In
http://www.unicode.org/Public//3.0-Update1/UnicodeData-3.0.1.html

Mn Mark, Non-Spacing
Mc Mark, Spacing Combining

1) The problem is, the entry reads 'Mn' -- non-spacing.
The way your examples use it, I would expect 'Mc' here.
The dot is plainly spaced to the right of the o, and takes up space.

2) Could there be different styles of usage of this letter?
Please check that. If that is the case, maybe I can do something.

3) It may be the Unicode standard is just wrong.
It wouldn't be the first time.)

4) I may be able to do something tricky anyway.
I'm not going to go against Unicode. But there are other options.

Mostly, I need to understand why the standard says one thing, and your images say something else, before I proceed.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 30 May 2015 10:03:33 AM UTC, comment #16:

Hi Hsing-yen,

I am transferring some of the files in your link to this site, for the record.

(file #34124)

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 30 May 2015 05:46:05 AM UTC, comment #15:

Hi,

The problem is not the font, at this point.

The problem is that your images are inconsistent with the Unicode standard.

Your images show a dot lower than the letter o and far to the right of its right side -- were another letter to follow, the dot would be inside the following letter.

What is shown in your images is not a combining mark, this is a spacing character.

But the Unicode standard clearly specifice U+0358 is a non-spacing, combining mark, which appears over the letter it is applied to.

We have a discrepancy.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Fri 29 May 2015 04:35:31 AM UTC, comment #14:

Steve,

I'm sure U+0358 is the correct character in this usage, and Pe̍h-ōe-jī is a romanization for Minnan dialect. In POJ orthography this dot is used for distinguishing vowels, as caron in Vietnamese. So I suggest U+0358 should be lower that near to O, as FreeSerif Regular's behavior.

Hsing-yen He <hammerandsickle>
Thu 28 May 2015 07:25:51 PM UTC, comment #13:

Hsing-yen,

As I look at your examples, I see something is wrong.

In these images, the dot appears far below the top of the o, and it is spaced far to the right of it -- about half the width of the letter o.

What appears in these images is a spacing mark, not a combining mark. It is impossible for a combining mark to appear that low and that far to the right of its letter.

Are you sure U+0358 is the correct character for this?
The Unicode description of the character is explicit: this is not a spacing mark.
And it says the mark is for "Latin transliterations of the Southern Min dialects of Chinese". Is Pe̍h-ōe-jī a "Southern Min dialect of Chinese"?

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Thu 28 May 2015 07:40:13 AM UTC, comment #12:

Well done, but one issue is: In Taiwan, this dot is lower than other diacritics in many books, magazines and newspapers using Pe̍h-ōe-jī, as shown in FreeSerif Regular. But this dot looks highter in other typefaces. Could you please fix it?

See:
https://commons.wikimedia.org/wiki/Category:Books_from_Taiwan
https://bible.fhl.net/new/ob.php?book=54&page=8

Hsing-yen He <hammerandsickle>
Mon 25 May 2015 01:58:26 AM UTC, comment #11:

Now I got newer revision to test, I found Serif Bold and Bold Italic performanced better than previous revision.

(file #34088)

Hsing-yen He <hammerandsickle>
Sun 24 May 2015 08:53:01 PM UTC, comment #10:

Hi Hsing-yen,

Sorry, I don't understand your last post.

Do you mean, you see no problems with the positioning,
or do you mean you don't think the problem is on your side?

If you are still seeing positioning problems (outside Serif Bold), please make a screen shot.

Thanks!

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 24 May 2015 06:21:53 PM UTC, comment #9:

Well, I have checked again, there is no problem for me.

Hsing-yen He <hammerandsickle>
Sun 24 May 2015 05:15:18 PM UTC, comment #8:

Please re-check your installation.

In the last build, I saw misplaced marks only in Serif Bold.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sun 24 May 2015 03:14:46 PM UTC, comment #7:

Hi,

It seems other 10 typefaces faild, I hope they are also fixed.

Hsing-yen He <hammerandsickle>
Sat 23 May 2015 03:16:22 PM UTC, comment #6:

Hi,

It looks like I failed to do the manual positioning in Serif Bold.
That will be fixed.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 23 May 2015 10:55:20 AM UTC, comment #5:

Hi Steve,

I got this build to test, I found FreeSerif Regular is rendering best, then Bold and Bold Italic looks well, but right dot looks farther, other faces looks worst.

Hsing-yen He <hammerandsickle>
Thu 21 May 2015 04:05:09 PM UTC, comment #4:

Hi Hsing-Yen,

Try your test with today's build.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Thu 21 May 2015 08:38:13 AM UTC, comment #3:

Sorry, I made a mistake in my attachment, now I think they should be fixed except O-carons.

Hsing-yen He <hammerandsickle>
Thu 21 May 2015 07:16:28 AM UTC, comment #2:

Hi,

I'm not sure what you're saying. Do you mean there is no need for special placement on the O-caron?

I see the dot isn't positioned well on O-caron. This is surely the same problem as we're seeing in your other bug reports. Something is mysteriously pre-composing O+caron to O-caron.

The problem is, it is a big mess, to put every positioning anchor on every character. There are thousands of Latin characters alone, and hundreds of posssible marks, and then... how many marks-on-marks? Before, we just positioned one mark based on a previous mark. Now...? I don't know what's going on.

Anyway, a policy will be: to place positioning marks where they seem to be required by typographic convention.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Thu 21 May 2015 05:31:18 AM UTC, comment #1:

Oh~sorry, POJ have no usage of caron when I reviwed the attached file.

Hsing-yen He <hammerandsickle>
Thu 21 May 2015 05:21:18 AM UTC, original submission:

Hi Steve,

I made O with right dot and other diacritics to test, many of them looks incorect, especially in non-normal typefaces.

Hsing-yen He <hammerandsickle>

 

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

Attach File(s):
   
   
Comment:
   

Attached Files
file #34134:  o-with-dot.jpg added by Stevan_White (10KiB - image/jpeg)
file #34124:  Ang-phoe_Seng-keng_te-571-bin.jpg added by Stevan_White (194KiB - image/jpeg - Shows dot mark after o)
file #34088:  POJ_test_result.zip added by hammerandsickle (587KiB - application/zip - These shots are based on revision 3058 and file #34083.)
file #34071:  Tests.zip added by hammerandsickle (546KiB - application/zip)

 

Depends on the following items: None found

Items that depend on this one: None found

 

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

    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 15 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Sat 29 Aug 2015 10:41:08 AM UTCStevan_WhiteAttached File-=>Added A_test_for_Pe̍h-ōe-jī_letter_rending.htm, #34763
      StatusNeed info=>Fix posted
    Sun 31 May 2015 07:20:31 AM UTCStevan_WhiteAttached File-=>Added o-with-dot.jpg, #34134
    Sat 30 May 2015 06:07:21 PM UTCStevan_WhiteStatusFix posted=>Need info
    Sat 30 May 2015 10:10:07 AM UTCStevan_WhiteOpen/ClosedClosed=>Open
    Sat 30 May 2015 10:09:13 AM UTCStevan_WhiteAttached File-=>Added Bible_in_Taiwanese.jpg, #34125
    Sat 30 May 2015 10:03:33 AM UTCStevan_WhiteAttached File-=>Added Ang-phoe_Seng-keng_te-571-bin.jpg, #34124
      Categoryoverall font problem=>individual character(s)
    Mon 25 May 2015 01:58:26 AM UTChammerandsickleAttached File-=>Added POJ_test_result.zip, #34088
    Sat 23 May 2015 10:56:44 AM UTChammerandsickleAttached File-=>Added A_test_for_Pe̍h-ōe-jī_letter_rending.htm, #34083
    Thu 21 May 2015 04:05:09 PM UTCStevan_WhiteStatusNeed info=>Fix posted
    Thu 21 May 2015 07:16:28 AM UTCStevan_WhiteStatusNone=>Need info
      Assigned toNone=>Stevan_White
      Open/ClosedOpen=>Closed
    Thu 21 May 2015 05:21:18 AM UTChammerandsickleAttached File-=>Added Tests.zip, #34071

    Back to the top


    Powered by Savane 3.1-cleanup1