Hi.
Not really Colemak, but I hope someone might find it interesting.
I want to:
Basically design a font for the standard latin alphabet, but one that have no visual baggage.
It should be optimized for fast reading.
How? I suppose with letter orthogonality and strokeminimization.
n-gram optimization
Are anyone in knowledge of literature relating to this?
Yes, I am insane and do this for fun.
Not quite sure what exactly it is you mean? Shorthand?
*** Learn Colemak in 2–5 steps with Tarmak! ***
*** Check out my Big Bag of Keyboard Tricks for Win/Linux/TMK... ***
Not quite sure what exactly it is you mean? Shorthand?
I'm not an expert, but I think shorthand is primarily fast to write, but therefore not necessarily fast to read. It think it can even be quite difficult to decipher really fast shorthand afterwards.
Just a visual redesign of the existing 26 characters.
'e' is still there, but maybe it is represented by a single horizontal stroke.
Optimized for reading speed.
Here I need to research if this is even possible.
What do I know, maybe there are enough dedicated neuralcircuitry that there are no speed to be gained.
Complexity of characters would be in proportion to frequency of the character.
(not by design, but there are only so many simple patterns)
One would have to observe digram frequency and make common sequences easy to disambiguate.
This is strictly for fun.
To do something like this correctly one would need control groups and cogscientists.
I would have considered special characters for common digrams, but then you couldn't use it as a 1:1 replacement font.
I would have considered special characters for common digrams, but then you couldn't use it as a 1:1 replacement font.
In fact, Unicode makes that possible.
In the English alphabet, character to glyph is indeed (mostly) a one-to-one mapping, but that's not true for many other writing systems, eg. Indic scripts and other syllaberies.
Unicode fonts can compose ligatures or even entirely new glyphs for certain sequences of characters.
Forgive me for possibly being far out there, but I'm currently learning Chinese and this is somewhat related.
In Chinese, each character represents one syllable (which usually is a word). There is no alphabet (pinyin does not count), so instead characters are built by combining other simpler characters in a block formation to create a new idea. For example, the character for good is a combination of mother and child next to each other. The character for bowl is a combination of stone and turned or curved, which is itself a combination of pictographs of the moon and a bent over person, signifying "turning over while sleeping".
While it does take a great amount of time to memorize these characters, I feel that once you have a good grasp of them it is quite easy to read since you're basically looking at pictures rather than combining individual letters to create words.
What does this have to do with the latin alphabet? I don't feel that simplifying the latin letters for faster or easier reading will be helpful, but rather harmful. The reason why I believe Chinese is quite easy to read (and one reason why I'm learning traditional rather than simplified characters for now) is because they are so distinct from each other. The visual "noise" if you will helps your brain understand what your eyes are telling it faster than if all the letters were simple and looked alike.
see also: http://www.arcavia.com/kyle/Software/Pr … index.html. I'm pretty sure this project is supposed to be a joke.
Ghen: Interesting, didn't know about the support for ligatures, always figured it to be part of some language pack.
Jrick: Easy disambiguation of letters is of course a consideration.
I think I see your point, suppose I have to ambush a neuroscientist at the uni and have him explain the visual cortex :)
I think this has been done before? But I can't remember who that American writer was... Whitman? No. There was someone back in the day, made a new font and claimed it'd be better to write and read while simpler in form. Too fuzzy on the details though.
*** Learn Colemak in 2–5 steps with Tarmak! ***
*** Check out my Big Bag of Keyboard Tricks for Win/Linux/TMK... ***
