Fascinated by languages in general and writing systems in particular, especially the complex abugidas of South and Southeast Asia. Reads Unicode proposals for new writing systems for fun.
Fontstructing since | 25th March, 2015 |
Fontstructions | 8 shared, 1 staff pick |
Shared Glyphs | 500 |
Downloads | 414 downloads made of this designer’s work |
Comments Made | 18 |
My second attempt at a font for the Marchen script, a Brahmic script used in the Tibetan Bon tradition to write the extinct Zhang-zhung language. It can also be used to write Tibetan. It supposedly originated in the Zhang-zhung kingdom prior to the introduction of Tibetan Buddhism in the 7th century, but no texts from that time using the script are known.
Marchen was added to the Unicode standard in version 9 released in June 2016. This font uses an ad-hoc ASCII mapping though. It doesn't handle stacked consonants which makes it rather useless. :P
This design uses the same rounded corners and serifs as my 'Phags-pa and Zanabazar Square fonts. The letters come in three widths. Combined with the medial 'y' there's a total of four widths that the vowel diacritics need to accommodate, which is a manageable number.
Marchen is a Brahmic script used in the Tibetan Bon tradition to write the extinct Zhang-zhung language. It can also be used to write Tibetan. It supposedly originated in the Zhang-zhung kingdom prior to the introduction of Tibetan Buddhism in the 7th century, but no texts from that time using the script are known.
Marchen was added to the Unicode standard in version 9 released in June 2016. This font is modelled very closely on the example characters in the Unicode chart which were designed (I think) by Andrew West. Like my other fonts it's not a Unicode font though but uses an ad-hoc ASCII mapping. Marchen, like the Tibetan script, relies heavily on vertically stacked consonants. I could in theory create precomposed compound characters for the most common stacks, but managing that with an ad-hoc encoding would be a nightmare. This makes the font rather useless. :P
My second biggest problem was that I wanted the vowel diacritics to be the same width as the base letters. These come in four widths. Add the medial 'y' which attaches to the right side of a letter and it turns into a huge mess. I solved this by creating extra "bars" that can be used to extend the diacritics.
I got this crazy idea that I wanted to create a script that looked like it had been created by tiny pastel horses (you're allowed to laugh). My inspiration comes from the Burmese script and a few other closely related ones that all kind of look like hoof prints stringed together at different angles.
There was just one problem: FontStruct doesn't do these circular letters. So I used FontStruct to create a prototype with mostly 5×5 and 7×5 letters to make the most out of the rounded corners.
Once satisfied I set out to find a bunch of tools to create "real" fonts. Inkscape and Fontforge looked like the perfect combination for a cheapskate like me. Designing the letters in Inkscape was easy since they consist of a few standard elements combined at different angles. These elements in turn mostly consist of circle sectors merged together. Fontforge on the other hand is a constant source of frustration. The constant crashes are the least of my problems. Paths that are imported will contain several errors that need to be fixed and however hard I try I can't seem to adjust widths and bearings in a way that makes sense.
Currently the script takes about 90% of its letters straight from Burmese while the vowels are more closely inspired by Shan. The script is used to write English phonetically and works as a not-really-an-abugida. There's no inherent vowel and it thus works more like Tolkien's Tengwar. Initial consonants are written as conjuncts while final consonants are mostly written as a linear sequence. The twelve vowels of RP have been merged into ten. There are two sets of vowel diacritics: the regular ones based on Shan and the overly cutesy ones consisting of hearts, gems and celestial bodies.
Todo: Thicker letters. Larger diacritics and increased distance. My own set of letters using the same basic components – possibly featural but probably not. Proper metrics. Kerning. Ligatures. Learn AAT and/or Graphite.
My second attempt at the Kayah Li script used to write Kayah Li spoken in Burma/Myanmar and also in Thailand. It was created by Htae Bu Phae in 1962. I used a smaller grid to be able to give the letters a more rounded shape. The 5×7 grid made some letters a pain to design though – especially the letter ‹s› which went through tons of attempts to even produce something that looked tolerable. I really like the thinner lines in ‹k›, ‹h› and ‹m› but they look a bit out of place. I tried making all diagonal lines in the font thinner but it just looked weird. There are a good amount of alternate glyphs hidden in the font in case I change my mind.
Kayah Li was added to the Unicode standard in version 5.1 in 2008. This font however uses an ad-hoc mapping to Ascii characters. The only real oddity are the tone markers mapped to 'f', 'j' and 'q'.
The script is a true alphabet with all vowels written out. There are however only four vowel letters: ‹a›, ‹oe›, ‹i› and ‹oo›. The rest are written as ‹a› plus a diacritic.
This is a clone of Kayah Li v1My first attempt at the Kayah Li script used to write Kayah Li spoken in Burma/Myanmar and also in Thailand. It was created by Htae Bu Phae in 1962. I based my design on the relatively blocky font used over on Omniglot. I'm not that happy about the result though. Letters were made on a 9×11 grid which allowed me to make vertical lines thicker than horizontal. It makes for very blocky letters though. I was hoping to create something similar to my Tai Le font but the letters simply didn't lend themselves to the same level of fluidity.
Kayah Li was added to the Unicode standard in version 5.1 in 2008. This font however uses an ad-hoc mapping to Ascii characters. The only real oddity are the tone markers mapped to 'f', 'j' and 'q'.
The script is a true alphabet with all vowels written out. There are however only four vowel letters: ‹a›, ‹oe›, ‹i› and ‹oo›. The rest are written as ‹a› plus a diacritic.
My take on the Mongolian 'Phags-pa script designed by the Tibetan monk Phagspa in 1269, based on the Tibetan script, to write Mongolian, Tibetan, Sanskrit and Chinese. This font is based on the Tibetan style which consists almost entirely of straight lines and right angles. It seemed like a prime candidate for a FontStruct treatment. I've added rounded corners and serifs to make it more visually interesting.
The script is written in vertical columns top-to-bottom and left-to-right and thus needs to be rotated 90° clockwise and the columns then reversed.
'Phags-pa was added to the Unicode standard in version 5.0 in 2006. This font however uses an ad-hoc mapping to Ascii characters which admittedly doesn't always make sense. I kind of gave up in the end and started assigning a bunch of letters to digits. Letters are connected into syllable block by a thin line (mapped to '-'), usually on the right-hand side. A straight line clashed wth the serifs so I made it into a small arch.
The script is an abugida: the vowel ‹a› is inherent in each syllable and thus not written.
My take on the Tai Le/Tai Nüa/Dehong Dai script which is used mainly in the Dehong region in southwest China. The relative blockiness of the letters made it a prime candidate for a FontStruct treatment.
Tai Le was added to the Unicode standard in version 4.0 in 2003. This font however uses an ad-hoc mapping to Ascii characters. Thanks to the limited number of letters (for a Brahmic script) the mapping mostly makes sense. Aspirated plosives are mapped to upper case and tone markers to shift+digit. The latter have been mapped so that they work on both US and Swedish Mac keyboards (and hopefully many others). Luckily there were no conflict between the two.
The script is an abugida: a syllable-initial consonant letter has an inherent vowel ‹a›. Whether a consonant is initial or final has to be inferred from context, however only ‹p›, ‹t›, ‹k›, ‹m›, ‹n› and ‹ng› can appear in final position.
(The letter pair ‹tone 2› + ‹ka› could really use some kerning.)
My take on the Mongolian Horizontal Square script designed by Mongolian spiritual leader Zanabazar to write Mongolian, Tibetan and Sanskrit. It's based on the Tibetan script. The script consists mostly of straight lines and right angles and seemed like a prime candidate for a FontStruct treatment. I've added rounded corners and serifs to make it more visually interesting.
The script has been accepted by the Unicode Technical Committee for inclusion in a future version of the Unicode standard*. This font uses an ad-hoc mapping to Ascii characters: upper case for aspirated plosives, 'f' and 'q' for retroflex plosives and a lot of mappings that make even less sense as I started to run out of Latin letters. The mapping is based on Sanskrit and Tibetan; Mongolian uses some characters differently. However, the font does not do stacked consonants required by the two former.
The script is an abugida: the letter ‹a› is inherent in each consonant letter and the vowel is then modified using diacritics. Initial vowels are written with a special letter, mapped to 'A', that's wider than the rest and has its own set of diacritics, mapped to digits 0–9.
* http://www.unicode.org/L2/L2014/14024-zanabazar-square.pdf