-*- org -*-
e.g. multiple words, part of speech etc.
Right now syllabics aren’t very pretty.
Also, maybe remove `text-align: justify`? Are there non-space spaces that get stretched in the jan Pije text?
Ideally the toggle should hover at top of screen so one can quickly toggle it when one wants to (if this is possible, remove the alt-texts). Remember, though it should be usable on mobile – maybe the toggle could go (almost completely) transparent after a while (on mobile only?; when user has scrolled down) so that text underneath it can still be read?)
In Javascript, probably use the following to sort in Inuktitut order:
list.sort((a,b) => a.localeCompare(b, ‘iu_CA’))
A la http://theotherwebsite.com/tokipona/
jan Sonja’s word based script
The Maya inspired hieroglyphics.
Cuneiform inspired script. https://www.reddit.com/r/tokipona/comments/i71a3g/sitelen_ko_a_script_for_clay/
Japanese inspired script. https://www.reddit.com/r/tokipona/comments/jax1x2/sitelen_telo_v101_a_japaneseinspired_logographic/
https://jprogr.github.io/TokiPonaDictionary/- Syllabics should have same line thickness in raised symbols as in other symbols.
- Overdot should be centered above the dotted circle symbol.
- “Tutorial: Choosing the right fonts to show Cree syllabics on your website” [https://blog.mothertongues.org/tutorial-syllabics-fonts/]
- “Google Noto Fonts” [https://www.google.com/get/noto/]
After playing around with fonts and discovering that next to no fonts actually have support for Canadian Aboriginal syllabics, I googled the problem and found the “Choosing the right font to show Cree syllabics” above.
This led me to “Noto Sans” which is the only font I’ve been able to find which have the same line thickness for raised symbols as for the other ones. For the Latin alphabet I tried “Noto Serif”, but the serifs look inane and unwieldy so a decided to use “Noto Sans” for all text.
BUUT it turns out that the placement of the overdot is not centered above the dotted circle in “Noto Sans Symbols”, so I instead went with Times New Roman from Windows 10 (which fixes this, and also have an overdot which have about the same size and shape as the overdots in the syllabics in “Noto Sans Canadian Aboriginal”).
In the end, “Noto Sans” is used for all Latin characters, using “Noto Sans Canadian Aboriginal” as a fallback to handle all syllabics symbols, and “Times” as an additional fallback to handle symbols (like the overdot, and dotted circle).
- Use ᐧ ‘final middle dot’ (inspired by Shavian) – if so, this requires not using dotted forms for the w-series.
- Use ᑊ ‘West-Cree P’.
- Use * (asterisk, borrowed from linguist’s marking of a word that is not correctly formed) – but this is less easy to write with a pen.
- Use ᐤ ‘final ring’
Marcus Miles wrote on facebook “I can hazard a guess at the o/u merger. I discovered long ago that if you merge o and u, you don’t get any accidental homophones.” (21 December 2019, 23:36, in Facebook group “toki pona”) [https://www.facebook.com/groups/sitelen/permalink/2695902890464079/?comment_id=2695938253793876&reply_comment_id=2696157930438575]
Zoltán Gorza 21 December 2019 · Public
I am working out a new writing system for toki pona called “sitelen lili”, and I’m already working on a font for it, but that might take a while. It basically a syllabic system, consonant (or “blank character” if there’s no consonant) is read first than the vowel above it, has multiple forms for each letter (for stylistic use or differentiating/emphasizing), has its own punctuation and o-u difference is optional. The long text is the first verse of “olin li tawa jan ale”.
[pictures]
Robin Morton
Interesting. I like its efficiency in terms of strokes.
A couple of questions. What made you decide to design the alternative forms? And the o/u optional merger? These seem to add a bit of complexity that I’m not sure is necessary.
I wonder whether there’s another stage you could push this to, where the alternate forms have usage patterns that make ergonomic sense, like the tilde-like n is used after a horizontal stroke to differentiate it, whilst the rounded forms of w and m might be used when sequences of your various ‘v’ / ‘L’ shapes get a bit tricky to distinguish.
Have you tried designing cursive forms for these? The angles are perhaps a bit confusing. Maybe look to Canadian Aboriginal Syllabics and/or Thaana and/or shorthand systems for some ideas about creating a flowing, yet distinguishable style with a bunch of pretty similar letter forms.
Hope you don’t mind my honest thoughts, fundamentally you’ve got a lot of good material there but if it were a script in real life I think it would naturally develop a bit past this point for elegance and ease 🙂
Like · Reply · Share · 1 y · Edited
Marcus Miles
Robin Morton: I can hazard a guess at the o/u merger. I discovered long ago that if you merge o and u, you don’t get any accidental homophones. The only consonant affected is < j >, And then only if you change o to u rather than u to o. So Toki Pona can function with a four vowel system rather than a five vowel system.
Like · Reply · Share · 1 y · Edited
Zoltán Gorza
Robin Morton if there wasn’t kin and ken, even i and e coupd be merged. as Marcus said, there are no words where only o-u differes, so it’s possible (and I haven’t had a good idea for ‘u’)
Like · Reply · Share · 1 y
Robin Morton
And tbf you could spell ken kan instead…
Like · Reply · Share · 1 y
For a word to be capitalized it must start with a capital letter, if it thereafter ends (being only one letter long), or contain at least one lower case letter then it is counted as a Capitalized Word. If the word is FULLY CAPITALIZED it will not get a naming mark, but if it is Boaty McBoatface it will. All words capitalized thusly get a naming mark (even if there are more than one after each other).