Ideas for a "bi-syllabic" orthographic theory.
161 views
Skip to first unread message
ReneSac
May 17, 2015, 8:42:14 PM5/17/15
to plove...@googlegroups.com
I was thinking an orthographic chording keyboard that could by default spill two syllables per stroke, if they are the normal CV (consonants+vowel) type. If not, you could usually construct a more complex CVC syllable, accentuated syllables and diphthongs using the two hands in one stroke. My thinking is that for chording to win against normal typing one needs to produce an average of at least 3 characters per stroke, but hopefully more. And in general computer/internet usage we usually have to type things that are not normal words, and in many languages, so flexibility is good. Thus an orthographic theory.
The worst case of finger spelling in my layout is two characters per stroke (for example, two consonants, two vowels, a consonant and a space, two symbols, etc), and the best case is ~8 characters in a completely regular stroke (can't think of a such a word now), before involving any auto-correction or briefs. Bear with me while I explain the theory. It is actually more of a "hypothesis", as I haven't tested it yet.
I was inspired by the GKOS chording keyboard and the USA steno keyboard layout (but symmetrical, more like the korean steno layout). Latter I saw that it resembles quite a bit the velotype keyboard, but still, is quite different. I haven't worked out the details yet, but here are the general points:
It is mostly symmetrical and mirrored between hands to facilitate learning and possible one handed use. Each hand is further split in 3 regions:
- The pinky, ring and middle fingers do a GKOS like keyboard with one hand. Two keys for each finger that can be vertically chorded like the USA steno keyboard.
- The index finger has control over 4 keys that output 6 vowels: a, e, i, o, u and y (for english).
- The thumbs are not symmetrical. One side handles Shift and Brief keys. The other handles "s" (poor man's plural for orthographic) and Space keys. They can be chorded together in any combination and the result would be the expected combination of individual effects.
----------------------------------------------------------------------------
Diving more deeply, the GKOS part is mainly responsible for the consonants, that will precede the vowel typed with the index finger in each hand. They don't need to be typed together, you can type only a consonant or a vowel with one hand. The letters the left hand type precede the letters the right hand types.
But that covers only ~20 of the 63 possible combinations with the 6 GKOS keys. We can use some of those for the most common consonant clusters, like "ch", "rr", etc, (and non-consonant clusters too, like "qu") that will behave like normal consonants and appear before the index finger vowel. We can have about 20~25 of those.
The best set to chose depends on the language, and is kinda like a 3/5-hand brief, but I still consider them a fundamental part of the base orthographic layout as they are a fixed limited set and combined just as a normal single consonant. W/o them many basic syllables would take the whole hand to type, and several more complex syllables would need two whole strokes. Of course, some rarer combinations will still be problematic, and need two whole strokes, like "clam", as "cl" likely won't be one of the chosen consonant clusters, and one would need to type "c" in one hand, "la" in the other, and then "m" in a second stroke. But those should be rare, by definition.
The GKOS part don't need to be isolated producing consonants to be put before the vowels. Some of those combinations can be common consonants that appear in VC types of syllables. For example: r, l, n, m, s, t. And they would be combined with the index finger vowels to form "an", "or" etc. I'm not sure what to do when they are typed alone, w/o a vowel.
Abstracting further, we can have GKOS combinations for "basic punctuation". Then the index finger will select between ".", ",", ";', ":", "?" and "!" instead of inputting a vowel. Another example: "braces": (), [], {}. And so on. Some of those might make more sense as a whole hand brief, though.
Using that last type of chord, we can also have things that will combine with the other hand. For example, accentuation for the vowel on the syllable of the right hand (if pressed alone, it can print the symbol alone, like if double pressing a dead-key). Also, chords to shift the right hand into a numerical (will be described bellow) or functional layer. Lastly, chords for modifiers and combinations of modifiers, like "ctrl + shift" on the left hand and then the other key with the right hand. As "space" and "plural" are meaningless in this context, the thumb can be used to give access the numerical and function layer then.
We might now kill the symmetry between hands. The last paragraph chords can be exclusive to the left hand. The right hand could then use that chord space for common terminations, like "tion", "ught", "n't", etc (the first vowel may be typed by the left hand). This seems specially useful for portuguese. They are a like briefs, but are of a very limited number, will probably have completely non-logical chords, and are typed with only one hand to be mixed in the orthographic mode. Some common prefixes could also be added to the left hand.
----------
The "brief" key is for pure macros, precisely defined one by one in a dictionary. This means that the orthographic layer is pure and fixed, and I can type "qx" in one stroke w/o fear or triggering a brief like in velotype. Also means that this whole layout can be implemented in a keyboard micro-controller as a combination of many small fixed look-up tables plus a user defined larger dictionary.
The logical mnemonics for the briefs will be two syllables, but it will have less consonant keys to combine compared with a normal steno keyboard. I don't know if this is a win. a lose, or neutral. I haven't really tough in a brief theory with this, but an really elementary example in portuguese: "esta" would be typed in one stroke orthographically. "está" could be a brief using the same stroke but adding the brief key.
Shift key toggles the capitalization of the first letter of the word or brief. The space key puts a space after the stroke, even for briefs (maybe the reverse is better). The "s" key outputs an final "s" before a possible space in orthographic mode, but is free to use in brief mode, though you will probably want it indicating plural for words where this makes sense.
---------
Some details on the planned one hand numerical layer. The thumb would be responsible for space, enter and tab. The other 10 keys would each be a number, lined like this (as an example):
0 1 2 3 4
5 6 7 8 9
(the home position for the index finger is "16")
The numbers can be keyed individually, of course. Chording any two keys will result in the two numbers being print, first the smaller one, then the larger one, like "19", "78", "05". If you press a number together with a column using another finger, you add a 0 after the number. If you chord 2 columns, add two zeroes, and three columns add three zeroes. Example: "1 27" chord or "1 38" chord would print 10. "16 2 38 49" chord would print 2000. That is why the 0 key is in one of the worst positions.
There is some redundancy in this system that might be used for other purposes, but I think it is better left as is. Chording 3 or more numbers in a way not covered by the 0 rule, like "1 2 3", "1 27 3, can be used to print other symbols like "+", "=", etc. It may also be a good idea to have a chord to print "00" and "000" alone. Something like "16 27" and "16 27 38" respectively. I don't think the full redundancy is needed for this.
The position of the 5 is kinda bad on this first one, so a less logical but more optimal layout could be like this:
0 4 1 2 3
8 9 5 6 7
The worst case of finger spelling in my layout is two characters per stroke (for example, two consonants, two vowels, a consonant and a space, two symbols, etc), and the best case is ~8 characters in a completely regular stroke (can't think of a such a word now), before involving any auto-correction or briefs. Bear with me while I explain the theory. It is actually more of a "hypothesis", as I haven't tested it yet.
I was inspired by the GKOS chording keyboard and the USA steno keyboard layout (but symmetrical, more like the korean steno layout). Latter I saw that it resembles quite a bit the velotype keyboard, but still, is quite different. I haven't worked out the details yet, but here are the general points:
It is mostly symmetrical and mirrored between hands to facilitate learning and possible one handed use. Each hand is further split in 3 regions:
- The pinky, ring and middle fingers do a GKOS like keyboard with one hand. Two keys for each finger that can be vertically chorded like the USA steno keyboard.
- The index finger has control over 4 keys that output 6 vowels: a, e, i, o, u and y (for english).
- The thumbs are not symmetrical. One side handles Shift and Brief keys. The other handles "s" (poor man's plural for orthographic) and Space keys. They can be chorded together in any combination and the result would be the expected combination of individual effects.
----------------------------------------------------------------------------
Diving more deeply, the GKOS part is mainly responsible for the consonants, that will precede the vowel typed with the index finger in each hand. They don't need to be typed together, you can type only a consonant or a vowel with one hand. The letters the left hand type precede the letters the right hand types.
But that covers only ~20 of the 63 possible combinations with the 6 GKOS keys. We can use some of those for the most common consonant clusters, like "ch", "rr", etc, (and non-consonant clusters too, like "qu") that will behave like normal consonants and appear before the index finger vowel. We can have about 20~25 of those.
The best set to chose depends on the language, and is kinda like a 3/5-hand brief, but I still consider them a fundamental part of the base orthographic layout as they are a fixed limited set and combined just as a normal single consonant. W/o them many basic syllables would take the whole hand to type, and several more complex syllables would need two whole strokes. Of course, some rarer combinations will still be problematic, and need two whole strokes, like "clam", as "cl" likely won't be one of the chosen consonant clusters, and one would need to type "c" in one hand, "la" in the other, and then "m" in a second stroke. But those should be rare, by definition.
The GKOS part don't need to be isolated producing consonants to be put before the vowels. Some of those combinations can be common consonants that appear in VC types of syllables. For example: r, l, n, m, s, t. And they would be combined with the index finger vowels to form "an", "or" etc. I'm not sure what to do when they are typed alone, w/o a vowel.
Abstracting further, we can have GKOS combinations for "basic punctuation". Then the index finger will select between ".", ",", ";', ":", "?" and "!" instead of inputting a vowel. Another example: "braces": (), [], {}. And so on. Some of those might make more sense as a whole hand brief, though.
Using that last type of chord, we can also have things that will combine with the other hand. For example, accentuation for the vowel on the syllable of the right hand (if pressed alone, it can print the symbol alone, like if double pressing a dead-key). Also, chords to shift the right hand into a numerical (will be described bellow) or functional layer. Lastly, chords for modifiers and combinations of modifiers, like "ctrl + shift" on the left hand and then the other key with the right hand. As "space" and "plural" are meaningless in this context, the thumb can be used to give access the numerical and function layer then.
We might now kill the symmetry between hands. The last paragraph chords can be exclusive to the left hand. The right hand could then use that chord space for common terminations, like "tion", "ught", "n't", etc (the first vowel may be typed by the left hand). This seems specially useful for portuguese. They are a like briefs, but are of a very limited number, will probably have completely non-logical chords, and are typed with only one hand to be mixed in the orthographic mode. Some common prefixes could also be added to the left hand.
----------
The "brief" key is for pure macros, precisely defined one by one in a dictionary. This means that the orthographic layer is pure and fixed, and I can type "qx" in one stroke w/o fear or triggering a brief like in velotype. Also means that this whole layout can be implemented in a keyboard micro-controller as a combination of many small fixed look-up tables plus a user defined larger dictionary.
The logical mnemonics for the briefs will be two syllables, but it will have less consonant keys to combine compared with a normal steno keyboard. I don't know if this is a win. a lose, or neutral. I haven't really tough in a brief theory with this, but an really elementary example in portuguese: "esta" would be typed in one stroke orthographically. "está" could be a brief using the same stroke but adding the brief key.
Shift key toggles the capitalization of the first letter of the word or brief. The space key puts a space after the stroke, even for briefs (maybe the reverse is better). The "s" key outputs an final "s" before a possible space in orthographic mode, but is free to use in brief mode, though you will probably want it indicating plural for words where this makes sense.
---------
Some details on the planned one hand numerical layer. The thumb would be responsible for space, enter and tab. The other 10 keys would each be a number, lined like this (as an example):
0 1 2 3 4
5 6 7 8 9
(the home position for the index finger is "16")
The numbers can be keyed individually, of course. Chording any two keys will result in the two numbers being print, first the smaller one, then the larger one, like "19", "78", "05". If you press a number together with a column using another finger, you add a 0 after the number. If you chord 2 columns, add two zeroes, and three columns add three zeroes. Example: "1 27" chord or "1 38" chord would print 10. "16 2 38 49" chord would print 2000. That is why the 0 key is in one of the worst positions.
There is some redundancy in this system that might be used for other purposes, but I think it is better left as is. Chording 3 or more numbers in a way not covered by the 0 rule, like "1 2 3", "1 27 3, can be used to print other symbols like "+", "=", etc. It may also be a good idea to have a chord to print "00" and "000" alone. Something like "16 27" and "16 27 38" respectively. I don't think the full redundancy is needed for this.
The position of the 5 is kinda bad on this first one, so a less logical but more optimal layout could be like this:
0 4 1 2 3
8 9 5 6 7
ReneSac
May 17, 2015, 8:44:35 PM5/17/15
to plove...@googlegroups.com
I made this system thinking on portuguese and japanese, with it's simple and short syllables, but it actually might work even better for english. All 45 most common english words (according to a Gutenberg corpus) can be typed orthographically in one stroke w/o any briefs! And of course one can always define a brief to type any word or phrase in a single stroke.
This is just a sketch though that I was thinking for the past month. I haven't tried to assign keys, much less using it (I don't know if I will ever use it to be honest). But I decided to disclose it here, as it may be at least an inspiration for other people to build their theories, as I saw someone has already done for spanish in another thread. Also, this will hopefully prevent someone like ASETNIOP patenting those ideas (I doubt he will ever recover the cost of his patent, actually).
I fear that the relative load on each finger or the ergonomicity of the chords might be bad. I don't have any experience with steno or steno keyboards, so I'm only basing on what I've read. For the average stroke my system seems to need more keys pressed than the normal steno system, but the thumbs might be under/inneficiently-worked. They could have their role changed to the common consonants that follow vowels in syllables (or accentuation), making many more words typable orthographically in a single stroke (not that critical given briefs) and the keyboard more symmetric (but maybe a bit more complex IMHO). Shift can be moved to a separate chord, "s" will be kept, but now on both hands. The Space and Meta keys are more fundamental. They could be moved to a palm key like in velotype (and space changed in "no space"). But I'm not sure if palm keys work well.
More keys can also be added to the design if they reduce the layout and chording complexity, but otherwise the current number of keys seems sufficient. The pinky can handle one more key to the side. Thumbs also could theoretically handle 3 keys each. The index finger is very flexible, and could handle another row of keys. Or we could go for a velotype style keyboard, with much more keys, but this would significantly impact the chording rules above, as the home position and key shape would change too.
I also don't know how well the brain can process 2 syllables at a time for the orthographic mode. In other words, I don't know if a human can take advantage of the many ways to write shorter w/o briefs that this system provides. On the other hand,
at any appreciable speed of typing, even on QWERTY, we learn to type particular letter combinations and whole words by muscle memory, like when reading. Thinking letter by letter is too slow. It is almost like briefing. So I'm not sure how big of a problem is it. Maybe alter the design to allow alternating strokes on the sides. It would make a one-hand version easier to do, but would disallow briefs using only one side among other things.
Even though I wanted something easier to learn than normal steno, and easier to use for multiple languages at once, I don't know if this design really attain this. Especially the large number of 3/5 and one hand chords that one must use for really fast typing, many of which language specific. I would define the relevant learning curve as "time to reach 50WPM", because at that speed it is already comfortable to type with it and start gaining speed naturally. I still haven't calculated what is the expected theoretical upper bound of writing speed on this keyboard.
Feedback is welcome.
This is just a sketch though that I was thinking for the past month. I haven't tried to assign keys, much less using it (I don't know if I will ever use it to be honest). But I decided to disclose it here, as it may be at least an inspiration for other people to build their theories, as I saw someone has already done for spanish in another thread. Also, this will hopefully prevent someone like ASETNIOP patenting those ideas (I doubt he will ever recover the cost of his patent, actually).
I fear that the relative load on each finger or the ergonomicity of the chords might be bad. I don't have any experience with steno or steno keyboards, so I'm only basing on what I've read. For the average stroke my system seems to need more keys pressed than the normal steno system, but the thumbs might be under/inneficiently-worked. They could have their role changed to the common consonants that follow vowels in syllables (or accentuation), making many more words typable orthographically in a single stroke (not that critical given briefs) and the keyboard more symmetric (but maybe a bit more complex IMHO). Shift can be moved to a separate chord, "s" will be kept, but now on both hands. The Space and Meta keys are more fundamental. They could be moved to a palm key like in velotype (and space changed in "no space"). But I'm not sure if palm keys work well.
More keys can also be added to the design if they reduce the layout and chording complexity, but otherwise the current number of keys seems sufficient. The pinky can handle one more key to the side. Thumbs also could theoretically handle 3 keys each. The index finger is very flexible, and could handle another row of keys. Or we could go for a velotype style keyboard, with much more keys, but this would significantly impact the chording rules above, as the home position and key shape would change too.
I also don't know how well the brain can process 2 syllables at a time for the orthographic mode. In other words, I don't know if a human can take advantage of the many ways to write shorter w/o briefs that this system provides. On the other hand,
at any appreciable speed of typing, even on QWERTY, we learn to type particular letter combinations and whole words by muscle memory, like when reading. Thinking letter by letter is too slow. It is almost like briefing. So I'm not sure how big of a problem is it. Maybe alter the design to allow alternating strokes on the sides. It would make a one-hand version easier to do, but would disallow briefs using only one side among other things.
Even though I wanted something easier to learn than normal steno, and easier to use for multiple languages at once, I don't know if this design really attain this. Especially the large number of 3/5 and one hand chords that one must use for really fast typing, many of which language specific. I would define the relevant learning curve as "time to reach 50WPM", because at that speed it is already comfortable to type with it and start gaining speed naturally. I still haven't calculated what is the expected theoretical upper bound of writing speed on this keyboard.
Feedback is welcome.
Mirabai Knight
May 17, 2015, 8:45:57 PM5/17/15
to ploversteno
Have you looked into Velotype/Veyboard? It seems similar to what you're looking for here.
Mirabai Knight
May 17, 2015, 8:46:34 PM5/17/15
to ploversteno
Oh, sorry, didn't see that you mentioned Velotype in your post. Disregard. (':
Ben Tarkeshian
May 18, 2015, 1:09:37 AM5/18/15
to plove...@googlegroups.com
"Feedback is welcome."
Sure :)
"For the average stroke my system seems to need more keys pressed than
the normal steno system, but the thumbs might be under/inneficiently-worked."
There's only one way to find out :)
Somewhere, hidden away in the heavily-guarded vaults of steno companies...only available to 342nd degree steno masons after extensive inititiation ceremonies and years of close supervision...
I suspect there must be a magical software that lets you feed it an input text and encode it as a theory, and get back statistics on:
Somewhere, hidden away in the heavily-guarded vaults of steno companies...only available to 342nd degree steno masons after extensive inititiation ceremonies and years of close supervision...
I suspect there must be a magical software that lets you feed it an input text and encode it as a theory, and get back statistics on:
- total number of strokes
- average number of strokes
- finger-count, key-count, chord count, etc.
and details on these things per word / sentence / paragraph / total.
And then one can compare across theories / input texts / etc.
So one can compare both "efficiency" as well as "this theory uses X fingers a lot" or "this theory is left-hand heavy" etc.
I'm sure such a thing exists, steno theory-makers quote such statistics.
I haven't looked ... but haven't seen programs focussing solely as steno "encoders" i.e.
plain text -> steno
everyone seems to want to go the other way...except theory-makers who need to compare versus other theories.
decode:
verb. to break apart into a well-defined format; see "encode"
verb. to break apart into a well-defined format; see "encode"
encode: see "decode"
- "normal" non-chorded keyboard layout (qwerty, dvorak, etc.)
- various machine shorthand theories
- various handwritten shorthand theories
- chorded keyboard layouts
Not quite 1:1 comparisons on everything, but "number of strokes for a word" I believe applies to all 4 styles.
"Even though I wanted something easier to learn than normal steno, and easier to use for multiple languages at once, I don't know if this design really attain this."
There's only one way to find out :)
"I punch those numbers into my calculator, it makes a happy face" -- Cave Johnson
"The logical mnemonics for the briefs will be two syllables, but it will have less consonant keys to combine compared with a normal steno keyboard. I don't know if this is a win. a lose, or neutral."
If I was a billionaire investor, I would say "gimme hard data"
I wouldn't use statistics as the final say in anything, but might point you better towards what "works" -- it gives you feedback, at least.
If it was me, that would be my next step, getting a "testing framework" and infrastructure set up, to run experiments/simulations, so I can "prove" whether any particular change makes a difference with regards to "efficiency" or not.
I wouldn't use statistics as the final say in anything, but might point you better towards what "works" -- it gives you feedback, at least.
If it was me, that would be my next step, getting a "testing framework" and infrastructure set up, to run experiments/simulations, so I can "prove" whether any particular change makes a difference with regards to "efficiency" or not.
And then "can a human use this?" is still another story, but data for various input texts, can presumably help with this too. (e.g. on finger movements, how many times had to leave the "home row" or whatever equivalent ... can have both "negative" and "positive" scoring mechanisms).
Selection of apropos "input sample texts" I'm not sure of an "objective" mechanism...you could presumably select certain inputs and make any theory "win" if you know what each theory/layouts is good and/or bad at.
Just my 2c.
It sounds like you are doing such calculations already to some extent, even if only "by hand" and "informally"...
If it was me, I would want to automate such "experiments" ASAP, so I know right away if some small change "helps" or not...
There must be some constant measure of "progress" or not that you can compare against.
That is the direction I would be going, if it was me -- how can I set up various "objective" measures of "goodness" and "badness" to test against other layouts?
There must be some constant measure of "progress" or not that you can compare against.
That is the direction I would be going, if it was me -- how can I set up various "objective" measures of "goodness" and "badness" to test against other layouts?
For me, that means "find/write software to tell me" -- maybe you do this somewhat already?
Roslyn Sim
May 18, 2015, 7:20:26 AM5/18/15
to plove...@googlegroups.com
You talk about comparing various steno theories against each other. I found this document useful: http://www.ncra.org/files/MCMS/5E781983-A6C9-4EA3-BCDD-B33ED6998381.pdf
Rafael Quintana
May 22, 2015, 8:45:02 AM5/22/15
to plove...@googlegroups.com
Hi, Rene Sac, it´s Rafael Quintana, Mecatype inventor:
In your last post you wrote:
"Even though I wanted something easier to learn than normal steno, and easier to use for multiple languages at once, "
Although I developed my Mecatype theory for Spanish,(easy and quick to learn, no briefs, no dictionary, orthographic theory, only 21 keys, +/- 200 wpm, and other advantages ) I suggested to Paolo Paniago for using it for portuguese.
You asked for anyone a feedback to your post. And here it´s. I hope that we, the spanish and portuguese speakers can make a common front for developing our own theory with our own keyboard layout , taking advantage of a common charasteristic: our languages are nearly orthographics.
Reply all
Reply to author
Forward
0 new messages