Some do but as with RSI from keyboard/mouse usage it affects everyone differently. One thing they usually have in common is excessive repetition for example by constantly using shortcuts with CTRL. While practicing typing (both keyboard and gamepad) I noticed that highly repetitive lessons with lots of text like "jf jf jf fj fj fj jfjfjfj" cause much more strain than lessons with more realistic text. In the end it all comes down to responsible usage, nothing is completely safe.
You're right of course about removing the letter j from the most used letters. My strategy is to continue logging the inputs made with this system and adjust the mapping as frequencies change. If I indeed use the d-pad more often than j when using this system, the letter j will eventually be flagged for remapping. By making only one adjustment at a time the relearning effort is manageable.
As for diagonal inputs, I've actually done it and I'm still experimenting with it. I can accurately target 8 sectors when there is no dialing involved but with dialing I would stick to 4. I think a hybrid system that allows for 8 sectors but switches to 4 when I start rotating the stick would be the best of both.
I will give it a try! You might still have to use both hands since you can't reach as many keys on a gamepad single-handedly as you can on a chorded keyboard.
You do have to be careful and stop when symptoms occur especially in the beginning but I believe the issue can be avoided. The key is to stay relaxed and focus on minimizing mistakes rather than trying to type very fast from the get-go. Typing with this system is actually not that different from playing a console game.
That would be helpful in the very beginning but even with just a little practice the gestures become muscle memory and remembering the gestures a non-issue. I believe making the gestures predictable would have greatly decreased efficiency which is my priority here as my goal is to be able to use this system for work not just for the occasional text message.
Indeed, when entering combined inputs one by one starting with my off-hand my accuracy was higher and I found myself using that technique occasionally during practice, especially when I had made a mistake and needed to repeat an input. However, as my speed increased, eventually I tended to move the sticks simultaneously.
Interesting article! I agree that mnemonics don't help you much with long-term retention, that's why it makes sense to combine it with a spaced-repetition strategy such as rote learning. However, mnemonics help you translate information into a form that's easier to process for the brain. Even if you don't actively use mnemonic techniques when memorizing meaningless information, you will likely end up applying such techniques intuitively.
Great idea about finding or even generating rhyming phrases! If you clearly define the parts of speech used for encoding (probably nouns), I don't think it would cancel out the benefit.
You're right that the voiced and unvoiced pair ð and θ probably belong together, as is the case with other voiced and unvoiced pairs. This would also significantly reduce the disambiguities that need to be manually resolved in Step 3. I chose to put θ into category 8 though because intuitively it sounds more like f than t to me, possibly because I'm not a native speaker. I might change my mind about it if it turns out to be more practical to have them both in category 1.
I like your idea of using the major system as a form of steganography. It would require a deterministic encoding algorithm though. I might explore this idea in another post in the future.