kittea

How to write with the font:

I wanted the font to translate the direct phonetic translation of the Latin alphabet that is specified within the glyph document to Yongasabi’s alphabet, but the glyphs in Yongasabi are syllabograms, leading to there being ambiguity when translating into it from the Latin alphabet which, if there is a constant or consonants between two vowels, does not specify which syllable  the consonants belong to. 

The font Opashoo made for Yongasabi uses apostrophes as a separator to break up multiple syllable words with ambiguous cases, but I opted to use capitalization instead as it allowed for me to pull off a couple of tricks to reduce the amount of ligatures I needed to make. There are a few rules for how use capitalization in order to write the font below:

-The onset (consonants that are behind the vowel in a syllable) is always capitalized, in onsets with multiple letters (ch, ng, sh) both are capitalized.

-The coda (consonants ahead of the vowel) is always lowercase, multiple letter codas have both letters in lowercase.

-Short vowels (a, i, o) are Capitalized if the coda is present, [v] [Cv] [Vc] [CVc]

-u is capitalized if the onset is NOT present, [U] [Cu] [Uc] [Cuc]

-The a in ae is always capitalized, the e is capitalized if the coda is present, [Ae] [CAe] [AEc] [CAEc]

-the e in ei is capitalized if the coda is present, the i is capitalized if the coda is present and the onset is absent, [ei] [Cei] [EIc] [CEic]

Numbers directly work in the font without any hassle, thank you numbers.

Here’s a punctuation list! Refer to the glyph document for what they look like:

full stop - period | pause - apostrophe | text end - (ligature) period x2 gate brackets - parentheses | exclamation - exclamation mark       question - question mark | !? - (ligature) exclamation + question mark   arrow brackets - greater than / less than | long dash - underscore       ellipses - (ligature) period x3 | short dash - hyphen  quotes - double quotes / single quote

The font does not have glyphs for f, q, r, v, x, or z as they are not used in Yongasabi, any symbols not mentioned in the above section are not represented either. 

Inconsistencies with the glyph document:

-Middle line glyphs (b, d, h, l, n, s, y) in the coda are supposed to connect with the middle line in syllables with o even if the upper glyph is non-connecting (b, g, k, l, p, s, t, w)  I did not include this as it would increase the amount of needed ligatures.

-the glyphs in the coda of ae syllables are supposed to be the same size as the onset glyphs. I did not include this as it would increase the amount of needed ligatures.

-h in the onset of syllables with o is supposed to connect to middle line glyphs, I just forgot about this one.

How I made this thing:

I decided on the sizing first, every glyph was designed as lines upon a 16 by 24 unit grid, the lines were then thickened to be two units wide so the full glyph takes up a 18 by 26 unit grid (all grid measurements I say later are based on the non-thickened values). The size of the Fontforge canvas is 1000 pixels tall which turns out to be a hair over 260 mm so each grid space represents 10 mm. 

Before I started working on designing the glyphs I knew that I couldn’t make ligatures for every possible syllable combination as that would require 1938 characters to represent every possible combination (17x6x17 [cvc]+ 17x6 [cv]+ 6x17[vc]) but in Yongasabi the coda is almost always in the same place so I could just have it as a separate glyph that is imposed over the previous glyph (I did this by having the spacing of the coda glyphs being a single pixel and moving the glyph behind both spacing bars) meaning I only had to do the [cv] ligatures. For u I made the [vc] ligatures and imposed the onset from behind (I couldn’t do the same for ae because the [ca] ligatures would already combine them together and so I just had to push through with [cae] ligatures.

Because the [CV] ligatures for [cvc] glyphs only took up half of the space they should have, I made a separate set of [Cv] ligatures to use without the coda and used the lowercase vowels to make them distinct. The a in ae ligatures is always capitalized as to make it so [Ca] or [a] syllables wouldn’t combine with [ei]. Because vowels in [CEic], [EIc], [Cuc], and [Uc]  didn't need to combine with the onset I used the capitalization of the i and u to determine whether or not to add the null symbol in the upper slot. 

After looking at the glyph document for some time I decided that I needed every consonant in 6 x 6, 9 x 16, 12 x 16, 15 x 16, 21 x 16, and 24 x 16 in order to create all the needed combinations. Because two 9 tall glyphs left just enough space for a bar for the [COc] and [CIc] glyphs with non connecting consonants (this means that the coda glyphs have to be 9 x 16), 12 tall glyphs were needed for connecting [COc] and [CIc] glyphs along with non connecting [CAc]

With 15 x 16 being used for connecting [CAc] glyphs. 21 x 16 is for [Ci] and [Co], 24 x 16 is for [Ca], and 6 x 6 is for all of the tiny glyphs used in ae, ei, and u.

 I modeled the glyphs in Autodesk Inventor with the 6 x 6 consonants, vowels, numbers, and punctuation were made individually. The 16 wide glyphs were all made together with the intention of being able to modulate all of their heights with the same dimension, with a separate extrusion that I could suppress in order to get the center line glyphs without their top line. These dxfs were converted into svgs, organized, and exported in their final forms (as well as being scaled up by 10 because I didn’t want to put all those zeros in when modeling them) in Inkscape. 

Because I had already made and labeled all of the needed glyphs in my files Fontforge was mostly just putting everything in its right place, setting the spacing for the glyphs (I put gaps of 70 pixels between everything and because the 180 x 260 mm glyphs are about 680 pixels wide most things take up 750 pixels), and making all the ligatures work properly. 

Miscellaneous things:

Because c and e are used in ligatures but not on their own I just made them the null symbol in their respective roles, putting Ac after a capital consonant creates the standalone consonant glyph shown in the glyph document.

The non connecting onset glyphs are a bit higher than the connecting ones so that they don’t connect in [Cuc] combinations but this means that they’re higher than the others in [Cu] and [CEic] combinations aswell.

Shoutout to Michael Harmon for making this series of Fontforge tutorials which have been nearly the only source of information I have used to make both this font and my personal conlang’s font.

The dress took a bit for me to really get down but in the end it's basically just a rectangle with two parallel long strips on the sides (I also really understand why three signals can't tie it without help I struggled a lot with tying this thing nicely)

Some pictures of just the doll, I wanted the files to print support free so the larger body parts print in two halves with little connectors I could push in; this resulted in lines but I really quite like the way they look on him :]

Finally the two test models I made (along with a pen for scale) The first was made before before I modeled the limbs and was just to check if the size felt right and I didn't accidentally make assembly impossible (fortunately everything worked!) The differences between the second model and the finial are minute but important such as improved shoulder articulation, hip joints that assemble better, and functional eyelids.