Issues in Noto Sans Myanmar for "Shan Language"

[SZTSIIT]

Title

Issues in Noto Sans Myanmar for "Shan Language"

Font

'NotoSansMyanmar'

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/tree/master/hinted/NotoSansMyanmar Date: 2020-08-09

Font Version

V 2.001

OS name and version

IOS 13 and Above Windows 10 Systemwide Mac OS Systemwide Linux Systemwide

Application name and version

It is Systemwide

Issue

Two issues in "NotoSansMyanmar" font for Shan Language. 1- Wrong Character Designs 1.1 Shan Consonant - ၼ For Shan consonant - ၼ, it should be one stroke instead of two. 1.2 Shan Consonant - ၽ For Shan consonant- ၽ, it should be one stroke instead of two too. 1.3 Shan Consonant - ၾ For Shan consonant - ၾ, it should be the same with consonant “ၽ” as above. 1.3 Shan Consonant - ႁ For Shan consonant- ႁ, it is shown as below. In "NotoSansMyanmar" font right now for this consonant “ႁ”, it is just the combination of Myanmar(Burmese)’s “ဂ” and “ှ”. As it is confusing and the way Shan consonant “ႁ” in “NotoSansMyanmar” does not exist in Shan Consonants.

2- Positioning Error When combining Shan consonants such as “ၾ”,” ႁ” with “ူ” sign, it is not the way it should be.

2.1 – Positioning for “ူ” sign As shown above in"NotoSansMyanmar", the way of sign “ူ” positioned outside of the consonant is totally wrong. It must be under the consonant. It is the same with Shan Consonant “ႁ” it should be under its consonant.

2.2 - Positioning for “ွ” sign When Shan consonants “ၾ” and “ႁ” combined with “ွ” sign

2.3 - Positioning for “ႂ” sign When Shan consonants “ၾ” and “ႁ” combined with “ႂ” sign.

In addition, the proper size for Shan Digits is way too big. When it is using with Shan words, it is unreadable. Please see my screenshot below.

Character data

Myanmar Unicode Map: https://unicode.org/charts/PDF/U1000.pdf Unicode Character Codes "ၼ" 107C "ၽ" 107D “ၾ” 107E ” ႁ” 1081 “ႂ” 1082 “ွ” 103D “ူ” 1030 Shan Digits

Shan Language Website

https://shn.wikipedia.org

Thank You

[marekjez86]

@SZTSIIT : thank you for a very thorough write up

[nizarsq]

@SZTSIIT Thank you for reporting this bug. I looked at the issues using latest NotoSansMyanmar. I can reproduce most of the issues with the following exceptions: 1.1 - Wrong Character Designs Shan Consonant ၼ. I see there is one stroke 2.2 - I can't reproduce the issue with positioning for “ွ” sign when consonants “ႁ” combined with “ွ” sign. But I can reproduce the issue when Shan consonants “ၾ” 2.3 - I can't reproduce the issue with positioning for “ႂ” sign when Shan consonants “ႁ” combined with “ႂ” sign. But I can reproduce the issue with Shan consonants “ၾ”

I have no experience with Shan language so I could be wrong @marekjez86 for second check. I used the fonts in the following directories: https://github.com/googlefonts/noto-fonts/tree/master/phaseIII_only/hinted/ttf/NotoSansMyanmar/NotoSansMyanmar-Regular.ttf https://github.com/googlefonts/noto-fonts/tree/master/phaseIII_only/hinted/ttf/NotoSansMyanmarUI/NotoSansMyanmarUI-Regular.ttf https://github.com/googlefonts/noto-fonts/tree/master/phaseIII_only/hinted/ttf/NotoSerifMyanmar/NotoSerifMyanmar-Regular.ttf

I also looked at NotoSerifMyanmar and it seem to be OK.

[SZTSIIT]

@nizarsq The Shan Consonant "ၾ", "ၽ" are still should be one stroke . I attached another sample with"Great Hor Kham Yangon" font and the Shan Digits are should be the same level with consonants. What you need to fix these issues? Please let me know. Do you need "Great Hor Kham" font to check with it ? Here the link for "Great Hor Kham Yangon" Font - https://bit.ly/32b40Uj

Thank you so much for quick jump on it.

[TServicmm]

May I suggest one things, When I check in Great Hor Kham Yangon the Shan alphabet style is close enough for Shan language. It similar like in NotoSerif Myanmar font style.

References: However, when I check with one native Shan speaking like Ominglot alphabet style. Just one place need to be fix. Also same like in Padauk font, it pretty close correct for Shan alphabet font style.

Anyway, I will provide the screen shot in few day for how other native Shan people view. @SZTSIIT & @nizarsq : your provided information also very helpful for this topic.

Thank you so much all of your technical provide.

[SZTSIIT]

@TServicmm Of course I am a native speaker. Thank you please provide your screenshot. Here how we write Shan alphabets.

[TServicmm]

Hello @SZTSIIT , That is so good you are also Shan native speaker. But, I am not native speaker just I help for them. I will also include your screenshot to my Shan native friends too.

Thank you.

[nizarsq]

@SZTSIIT As I mention in my first comment I was able to reproduce most of the issues you reported with exceptions to what I mentioned earlier. BTW what Apps are using to test the font? Thanks

[SZTSIIT]

@nizarsq So the consonants "ၾ" and "ၾ" you can't make one stroke? There is already one consonant similar to these two is "ꧤ" (A9E4) http://unicode.org/charts/PDF/UA9E0.pdf. May be you just look at it and fix it. I don't know how google making the font for design.

I used MS word and text editor to test it. The below is the Pengram to test Shan font.

Pengram for Shan Language (1) Regular Shan Language- ပေႃးၶႂ်ႈၵတ်ႉၶႅၼ်ႇ ၵျႃႉလႅၼ်ႇၵႂႃႇၸိုင် တေလႆႈႁွင်ႉသြႃႇလီ ရဵၼ်းႁူႉသွၼ်ၽိုၵ်း ၺႃးမိူဝ်ႈၾပၢင်ႇႁၢင်ႇႁၢႆး ယိင်းလႄႈၸၢႆးဢိူၺ်း ပေၼိုၵ်း မႃးတႃႉ သိူဝ်းၵျေႃႇ တိုၵ်းထႃႈၶၢဝ်းၸမ်ၶၢဝ်းၵႆၼၼ်ႉသႃႊ။ ၼႆႉသမ်ႉပဵၼ် တႃႇၽိုၵ်း ပႃႇလိတႆး တူဝ်မူၼ်းဢေႃႈ

(2) Shan Pali- မိူဝ်ႈၼႆႉ ၽွင်းသင်ႇꧠႃႇရတၼႃႇ ႁေႃးတဵတ်ႈပꩦ်ႉꩧၢၼ်းတြႃး ၺႃးၸဝ်ႈၻႃႇၼ​ေꧤႃးၷ ၿလ ꩩၼ ႁႅင်းယႂ်ႇၼံထႃး ၷုꧣပူႇꩡႃႇ ဢႃႇယုဝꩨ်ႉꩩၼလႄႈ ၶူဝ်းၶွင်ၸေးပႃး ၾီလူးၵျေႃႇဢႃႇꩮႃႇဝၵ ၸိုင်ႈမꧡ်ꧢိမ လႆႈႁုၵႂႃႇၵျႃႉၼႃႇ ပေႃးႁွင်ႉသႃႇထုသၢမ်ၶႃးယဝ်ႉသႃႊ။

Thank You

[SZTSIIT]

@nizarsq I don't know what software you have to use to create but my possible solution is to write another "Ligature" code to point out ၾ+ ူ to correct designed block for ၾူ. So have to create another block for it. Do you want me to make a change and give you the font?

Thank you

[TServicmm]

After I conversation and discussion with my Shan native friends, they give language sample from the book. In our discussion, before in Shan language they use(on printed book) and write base on Shan language standard rule.

Modern time, some people create the font/keyboard and the design are not base on Language Standard rule. Each website are used differently font style. So, the people are confused using on their language. That is one example for Shan, I notice not only for Shan but also some languages same problem too.

As for us, we recommend to use from this below attachments text style and design for to get more accuracy language rule.

@SZTSIIT you can continuous your forum. We give just suggestion for to consider.

Thanks everyone again.

[SZTSIIT]

@TServicmm Thank you I see that the consonants ၽ and ၾ in your suggestion are not as it is. Because when we write on paper in reality, you can read it though. I will share it with other Shan expert anyway. Thank you. Can you PM me here I want to discuss with you for my other projects.

[SZTSIIT]

@marekjez86 Since your tech having issue to fix the issues will you allow me to fix it? I will fix the design error and add necessary code in it then send it to you to check it. If so can you give me the lasted version that @nizarsq fixed ?

[TServicmm]

@TServicmm Sure. How may I help you? By the way, what is long term for PM? Thank you.

[SZTSIIT]

@nizarsq @marekjez86 I fixed the issues that I reported above and below is the screen shots of what it is look like. I adjusted the position and size for digit too.

[SZTSIIT]

@nizarsq Please don't forget my Shan issue :D when ever you have time.

[ohbendy]

It's great to have specific feedback from native Shan users, thank you.

I designed Noto Myanmar Serif, based on the resources I have, and in consultation with native Shan readers. I did not do the engineering of these fonts. Considerations from a type designer:

1. Shan Na ၼ uni107C. In handwriting, it is true that the loops of the two bowls would join, but when making fonts, things are often treated differently, more typographically. In the Shan materials I have, both ways are attested, and the widely-used Padauk font also follows the approach taken in Noto Sans Myanmar. The Windows system font Myanmar Text does the same. When I made Noto Serif Myanmar, I was informed that readers have a preference for the continuous construction, so that's why it is different. But we cannot say the other construction is wrong per se, here's a good example I found in Hsipaw last year.

2. Shan Pha ၽ uni107D. There are actually two related parts to this question, the top-right terminal of the left bowl, and the way the left bowl joins the loop of the right bowl. I'd agree that the left bowl should probably not have the spur on its top right side, though again, if we refer to Padauk and Myanmar Text, those do have the spur there. As for the continuous join between the left bowl and the loop, this would appear to be another stylistic choice: in the Shan books I have there is a preference for the continuous version, but both are found. Again, we don't always follow handwriting, and different fonts are allowed/supposed to treat things differently.

(The indentation of the left bowl can face to the left, or upwards, in case that is also a question.)

3. Shan Fa ၾ uni107E. Indeed the construction of the bowls always follows Pha ၽ. My own recommendation for the leg is to keep it slanted, so that it doesn't interfere with any other marks below, and that was my intention in Noto Serif Myanmar. Unfortunately it was engineered so that the -u and -uu vowels get substituted with post-base forms instead.

4. Shan Ha ႁ uni1081. Again, my personal recommendation is to follow the sloped leg, to avoid interference with other belowmarks. Though, looking at the 21 August image above, we see the vertical leg is also shown in a writing primer. The diagonal leg was preferred by my Shan consultant. The bowl is usually more open than in the letter ဂ.

5. Positioning of -u and -uu vowels ◌ု ◌ူ vs post-base forms. Generally the post-base forms are only used if there are other marks below a consonant. But of course if the consonant has a descender (like ဋ or ဠ) the post-base forms are preferable, especially in a bold weight where there's not enough room for below-base -u and -uu. Not all solutions that work for a light weight such as the sample font presented above can work for a style range as broad as Noto, which has very bold and very condensed styles. If the descenders of Fa ၾ and Ha ႁ are diagonal, we can fit the -u and -uu below without problems.

6. Shan Fa ၾ and Ha ႁ with -aw vowel ◌ွ (Myanmar consonant sign medial wa). I definitely agree that the circle should sit to the right of the descender. This may be a case for a precomposed glyph for these two combinations.

7. Shan Medial Wa ◌ႂ uni1082. Agree, same as notofonts/noto-fonts#6. It should definitely not sit below, as in Noto Sans Myanmar.

8. Size of the numerals. This is a good question, actually two questions: size and alignment. In most digital fonts the Shan numerals are the same size as the Burmese numerals, and they all descend below the baseline. The numerals are not as common to find as the letters, and the resources I have don't give a definitive answer either way. I would want to see more examples of the numerals sitting on the baseline, not just a single digital font. One possibility would be to include both styles in the same way we include lining and oldstyle numerals for Latin.

Remember also that different Shan speakers have different preferences. I'll make contact with some others in my network to find out their opinions.

[SZTSIIT]

@ohbendy First of all thank you Ben for enginneer these fonts including Shan Language. But that incorrect shape and construction is not acceptable as it is not follow our Shan standard. That is why I am reporting here not because of how you design the font. We have plenty of fonts with various designs and most of those are following Shan Standard. Here are some examples.

1, For Shan Na ၼ U107C is correct you do have to do anything with it. I already reported to Microsoft about the issues what are need to be corrected in "Myanmar Text " font. Shan consonants in both "Myanmar Text" and "Padauk" are incorrect. Thanks for visited to Hsipaw between did you had a chance visit to Shan literature and culture too?

2. Shan Pha ၽ U107D and Fa ၾ U107E : Please don't follow Padauk and Myanmar Text it is incorrect. It should not have spur on the top and it should be one stroke. Below is a good sample.

3. To keep the leg for Shan Fa it slanted? Please do. You are right it doesn't interfrere with any other marks below to keep -u and -uu below it you should move it leg to the left a bit. Here is how Great Horkham Font design it is-

4. Shan Ha it really need to get fix. Below is how Great Horkham Font designed it.

5. There is no such forms in Shan Language it must stay below a consonant. To get enough room as I mentioned above, move its leg to the left. Please see below how Great Horkham Font works.

6. Here I will just give you how Great Horkham Font works again.

7. It is different with Burmese. For shan digits it should above the line not below as it is in Burmese. Here the example-

Thank you again Ben . I totally agree with you Shan speakers have different preferences but the issues above are not the font design it is the Standard issues are those are not exist in Shan language. I am here to help if you need anything please reach out to me.

[ohbendy]

Shan consonants in both "Myanmar Text" and "Padauk" are incorrect.

Thanks. I'll bear that in mind. I was also looking at Pyidaungsu and Tharlon. Are there other fonts that handle Shan in a good way?

---

1. In your first image, it looks to me as though five of the ၼ glyphs do not join the loops. With plenty of other attestations showing the loops not joining, I think your comment 'ၼ U107C is correct you do have to do anything with it' must mean 'you don't have to do anything with it'?

2. and 3. A couple of examples from my Shan books:

Are you saying the ၽ and ၾ this was are 'wrong'?

4. If I understand correctly the ႁ can have the leg vertical (as in your example) or diagonal (as in my image directly above). Either way it must allow for other marks to sit in the descender space.

5. There is no such forms in Shan Language it must stay below a consonant. If you're talking about the post-base forms of -u and -uu, I've seen them used in Shan sequences such as ၵေုႈ and ၵျိူၵ်ႈ. I'm also not sure how ◌ႂ with -u or -uu would look if you're saying post-base forms don't occur in Shan.

6. Looks like a good solution to me.

7. We can certainly move the Shan digits upwards to sit on the baseline. Do you have other examples please?

---

Bonus question: are you familiar with the Shan Pali alphabet? One of my Shan books shows a Shan form of Great Sa:

[SZTSIIT]

@ohbendy I meant for certain words. For ၾ, ႁ consonants with legs like this - uu must stay below its consonant. Examples- ႁူဝ် ႁူ ၾူ ၾူဝ် ၾုဝ် . We do have those like ၵျိူၵ်ႈ. Sorry for not mention clearly. ႀ this is Pali သ In your book is wrong typo. Do you want Pali consonants the updated and corrected one? Here Shan Pali in this sheet are correct https://unicode.org/charts/PDF/U1000.pdf . Again Ben thank you so much.

[ohbendy]

Both သ and the Great Sa are used many times in the text, so they can't be the same character, see 'sambuddhassa',

More examples where they're used contrastively:

[SZTSIIT]

@ohbendy Your book is outdated. They printed with incompleted font and it has many mistake. Here Shan Pali converter this might give you a perfect reference. You will find all shan vowels and so on of course the font are still your noto sans font so . https://aksharamukha.appspot.com/describe/Shan

[ohbendy]

So you would represent double သ with သ္သ or ဿ in Shan Pali?

[ohbendy]

On the Aksharamukha site, clicking 'conjuncts' is showing all sequences with asat instead of virama. These should all stack, I presume?

It's fantastic to see all those conjuncts, I have only a couple of old Sanskrit conjunct charts so will add these to make sure we test all possibilities in the next font revisions. Is it your site?

[SZTSIIT]

@ohbendy It is not my site it is belong to some group who studied the language. Can you give me sometime? I will gather full details about Shan Pali and post back here.

[ohbendy]

That would be extremely helpful, I'm always seeking more resources so we can keep improving the fonts. Thank you very much.

[patchew]

AFAIK, Shan avoided stacking early on. Sai Kam Mong's book on Shan scripts tends to show images from some manuscripts where Pali forms used the asat and do not stack. The only counterexamples for stacking that I saw were explicitly Burmese titles and am unsure if it's normative or aberrant.

On Thu, Oct 1, 2020 at 5:53 AM Ben Mitchell notifications@github.com wrote:

On the Aksharamukha site, clicking 'conjuncts' is showing all sequences with asat instead of virama. These should all stack, I presume? [image: Screenshot 2020-10-01 at 13 51 43] https://user-images.githubusercontent.com/12471463/94811454-4aae4700-03ed-11eb-8884-3513d2f456cf.png

It's fantastic to see all those conjuncts, I have only a couple of old Sanskrit conjunct charts so will add these to make sure we test all possibilities in the next font revisions. Is it your site?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/googlefonts/noto-fonts/issues/1811#issuecomment-702113349, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABUJFLFKFNXFKASXO45MLOLSIR3TVANCNFSM4PZGXFAQ .

[SZTSIIT]

@ohbendy Most Shan monk they prefer not using stack form in modern Shan Pali Study. Here is an example- ၼ​​မေႃ တသ်သ ၽၵဝ​​တေႃ ဢရႁ​​တေႃ သမ်မႃ သမ် ၿုၻ် ꩪသ်သ. Yes you are right about Sa and Great Sa. To get more clear if you use GHK Taunggyi font with Aksharamukha.

Shan Pali Consonants in newer printed book.

Pyidaungsu and Tharlon font are good. Please check Great Horkham font too. Here the link for "Great Hor Kham Yangon" Font - https://bit.ly/32b40Uj . It is developed by my institute developer.

[ohbendy]

Thanks both! I'm aware there are quite a number of different orthographies for Shan, as discussed in Sai Kam Mong's book, and it seems that some people might want to use stacked conjuncts, I think. Since the fonts contain both asat and virama, users can decide whether or not to stack their conjuncts. I think it's most robust if we test stacked conjuncts, even if most people will not use them. I'll compile a list of all the Pali-Sanskrit ones contained in the lists I've examined, for Burmese and for Shan characters.

[ohbendy]

@SZTSIIT I noticed Great HorKham Yangon and Taunggyi include glyphs for subjoined ဧ. What language(s) would use that?

[ohbendy]

Ok here's the list of conjuncts I've compiled for Burmese: https://github.com/ohbendy/Myanmar-font-resources/blob/master/Burmese%20Sanskrit%20conjuncts.txt

Shan will follow later.

[SZTSIIT]

Thanks both! I'm aware there are quite a number of different orthographies for Shan, as discussed in Sai Kam Mong's book, and it seems that some people might want to use stacked conjuncts, I think. Since the fonts contain both asat and virama, users can decide whether or not to stack their conjuncts. I think it's most robust if we test stacked conjuncts, even if most people will not use them. I'll compile a list of all the Pali-Sanskrit ones contained in the lists I've examined, for Burmese and for Shan characters.

Thank You Ben. We will use it when we convert all Shan Books into digital format.

[SZTSIIT]

@SZTSIIT I noticed Great HorKham Yangon and Taunggyi include glyphs for subjoined ဧ. What language(s) would use that?

For this I don’t see the use in Both Burmese and Shan. It is a vowel and it is not stay bottom. I will look into it and May be some language using it. Thank You

[SZTSIIT]

Ok here's the list of conjuncts I've compiled for Burmese: https://github.com/ohbendy/Myanmar-font-resources/blob/master/Burmese%20Sanskrit%20conjuncts.txt

Shan will follow later.

Thank You.

[ohbendy]

I'm reviewing the Shan letterforms now, and in my resources found a bit of variation in the shape of Pha (107D) and Fa (107E). For the left bowl, is there any preference among Shan readers? The first seems more common in my books.

[SZTSIIT]

@ohbendy That is not our Bha it is support to be Bha which we use in Shan Pali. Please see my screenshot below. Thank You 🙏

[ohbendy]

Ah, so the one with the closed loop on the left bowl is actually Bha? Then Pha can be the shape of the first or third one?

[SZTSIIT]

@ohbendy Yes the one with closed loop on the left bowl is actually Bha. The shape is can use both the first one and third one as long as the connecting stroke between left bowl and right bowl is forming one stroke. But in real hand writing most of us use the third one. The following is how Unicode use is perfect.

[SZTSIIT]

Is this issue is closed per virtualvinodh? What was the replacement for? Please.