Mongolian variant forms

Other notes:

The glyph produced is the same as is produced for the suffix for all the fonts tested.

Notes from Greg Eck about m+fvs2: "The only place where the single tooth initial 1820 is used is with the masculine ablative suffix . Given the standard NNBSP model, this is an initial position and not a medial position. My thoughts here are similar to the 1836 final situation – the spec called for a medial. Font implementers created the 1820 medial+fvs. The display at standalone produced the desired effect – you could view the single tooth initial. Implementation went on to test in context. The implementation did not work. The contextual implementation was placed at initial position. The fvs stand-alone display was left at the medial position resulting in a mis-match of the stand-alone with the contextual implementation. If this is correct, then I suggest that the medial+fvs2 spec be dropped and the initial+fvs1 spec be added for the single tooth initial ."

"One more point that might be considered – if the ablative suffix ACA/ECE was indeed being processed in the medial position, then one would expect there to be a glyph for this at the 1821-E medial slot. There is none."

Bolorsoft lists two additional forms: f+fvs1 , used for "bi (meaning "I"), bui (meaning "is")", etc.; and f+fvs2 .

The Bolorsoft table used for the BS column lists one additional form: f+fvs2 , where the glyph is the same as for the ordinary final form but should have absolutely no NIRUGU, and which is used for genitive case suffix. The non-variant final form has to have a small "nirugu".

The Bolorsoft table used for the BS column lists three additional forms: i+fvs1 , and f+fvs2 and f+fvs3. Both the latter two look the same as the ordinary final,. The third form is reserved for "khalkha" form, and could be simply ignored.

The Bolorsoft table used for the BS column lists one additional form: m+fvs1, but it looks the same as the ordinary medial, .

The 3rd medial form in Standardised Variants becomes the 2nd final form in the 2014 proposal, however the latter shows the same glyph for 2nd and 3rd medial forms. It should show the slightly taller variant (used in Todo) as the 3rd medial form (shown as the 4th medial in the other docs). There is no 4th medial in the 2014 proposal.

The following from Greg Eck explains the how some forms in the Q column were interpreted:

Given the initial premise from my last email "The presentation form = the isolate form (or the first isolate in the case of a character with more than one isolate)"

#1 The MGWBM shows the male loop "ml qa initial form". This is the presentation form (and at the same time the first isolate).

#2 The MGWBM shows the loop without dots "ml qa feminine isolate form".

#3 The MGWBM shows the loop with dots "ml qa feminine isolate form with dots".

So, it is reasonable to assume that #2 above, the loop without the dots, is the primary as it is the first in the list. Not only that, but the main difference between 182C and 182D is that 182C does not carry the dots whereas 182D does carry the dots. Since one of the main characteristics of the QA is that it does not carry dots, you would think that the primary isolate would also be without dots. So, that would make #2, the "ml qa feminine isolate form", to be the second isolate. As the loop with the dots, #3, follows rather than precedes, it is reasonable to interpret the MGWBM as saying that the "ml qa feminine isolate form with dots" is the third isolate. You need to have both to represent the formation of ligatures between the following combinations:

Initial_182C + Medial_Feminine_Vowel
Initial_182C + Final_Feminine_Vowel
Medial_182C + Medial_Feminine_Vowel
Medial_182C + Final_Feminine_Vowel

As with the feminine consonant Todo 184D, however, the feminine loop forms of the 182C isolate (what I refer to above as the second isolate and the third isolate) will never appear on their own. The isolates are designed merely to "represent" the presence of their ligatured forms.

NOTE: #3 glyph is used only in archaic Mongolian literature

The Bolorsoft table used for the BS column lists one additional form: i+fvs2 .

The USVL form shown for f+fvs1 was incorrectly placed in the document by error, and should have appeared as f+fvs2. (It is left as currently shown above.)

The NP for final+fvs1 is the over-ride to allow for the rightward sweep when context dictates a leftward sweep. That is why it looks the same as the final character in the chart.

The following from Greg Eck explains the how some forms in the Q column were interpreted:

#4 The MGWBM shows the male loop "ml ga initial form with dots". This is the presentation form (and at the same time the first isolate).

#5 The MGWBM fails to show the loop with dots but is expected to be there as it is in 182C. I think this is merely an omission.

#6 The MGWBM shows the loop without dots "ml ga feminine isolate form".

As with 182C above, it is reasonable to assume that #5 above, the loop with the dots, is the primary even though it is not listed. Since one of the main characteristics of the GA is that it does carry the dots, you would think that the primary isolate would also be with dots. So, that would make #5 to be the second isolate. As 182D ligature behavior closely follows 182C above, it is reasonable to interpret the MGWBM as saying that #6, the "ml ga feminine isolate form", is the third isolate. You need to have both to represent the formation of ligatures between the following combinations:

Initial_182D + Medial_Feminine_Vowel
Initial_182D + Final_Feminine_Vowel
Medial_182D + Medial_Feminine_Vowel
Medial_182D + Final_Feminine_Vowel

As with the feminine consonant Todo 184D, however, the feminine loop forms of the 182D isolate (what I refer to above as the second isolate and the third isolate) will never appear on their own. The two isolates are designed merely to "represent" the presence of their ligatured forms.

NOTE: #6 glyph is used only in archaic Mongolian literature

The Bolorsoft table used for the BS column lists one additional form: i+fvs3 , but the shape is the same as m+fvs2.

Analysis for Professor Quejingzhabu's Mongolian Specification 9.0 and by Greg Eck suggest that the 2014 proposal is correct for the medial forms, and that therefore the fonts shown above will need to be changed.

Note from Greg Eck: "Medial defaults were changed in specification 9.0 of Professor Quejingzhabu's work. Baiti has implemented these changes. Probably FangZhen, MenkhSoft, and JadeBird have followed suit. Not sure as of present. Suggest we wait for data to come in from these three fonts on the 1836 implementation before evaluating further."

Another note, specifically about m+fvs2: "I question whether FZ, NS, and MS are really used/needed/fully-functional_in_context_of_OpenType_rulings and therefore may be able to be removed. If they were removed, then the question of a specification at 1836 Medial+FVS2 becomes moot. The reason I doubt that they are actually in use is that the same glyph also appears in the final position. The thesis for my argument is that the glyph only shows up before MVS/NNBSP. This position given the MVS/NNBSP model is a final position. It is not a medial position as determined by the standard MVS/NNBSP model. My thought is that the medial+FVS2 variant was implemented to follow spec. At implementation time, it was found that the spec was wrong and the implementation was needed at the final position. The final implementation was made. The font displayed correctly. The result was a primary contextual implementation at final with a secondary FVS implementation at medial. What I am suggesting is that the spec be changed to match the common implementation of Baiti, Almas, and most probably FZ, NS, and MS also. So, a good discussion would be healthy here. Such a spec change would knock out 1836 medial+FVS2 completely and leave 1836 final asis. That leaves the argument on the default forms for the medial / medial+fvs1 positions for a later discussion."

The 2014 proposal moves the 1st medial form to the 2nd medial form due to research by the Professor that indicates, that according to syllable-forming rules, the medial loop so long considered to be a vowel, UE (U+1826), was determined to be the semi-vowel WA (U+1838) instead.

The same goes for the 1st and 2nd final glyphs.

The Bolorsoft table used for the BS column lists all the shapes given for KA here under KHA, and vice versa. This is because Bolorsoft uses KA (for western words) and KHA (tibet and chinese words) the other way around than Unicode.

The Bolorsoft table used for the BS column lists all the shapes given for KA here under KHA, and vice versa. I'm guessing that this is just a mistake in the table, so I have left the shapes as initially reported, before the table became available.