#Layout Optimization with a novel M4G approach based on real typing data
31 messages · Page 1 of 1 (latest)
I’ve been waiting for this, thanks for taking the time to share your approach and thoughts. It was an interesting read and I think it shows once again how personal a lot of these decisions are, because there are quite some things I feel different about. Maybe I‘ll find the time to ask about some specific stuff later
I would be happy to hear your opinion! 🙂
This is great! I got my bedtime reading material for the next couple days. Thanks for sharing this!!
Alright, you lost me at section 2 sub 1. >.<
Still Reading though, It really picked up for me at section 6...
Haha ok no problem. What exactly are you stuck with? Maybe I need to explain it with a little more detail. Unigrams are basically how often single keys occur in a text. Bigramme the same for 2 key combinations and so on
the numbers you just started to spew out, out of nowwhere after defining the gram stuff...Litterary logic(?) is something I was trying to get a grasp of (im autistic and a slow learner)....as of getting the CC1, and this is getting my back into it
The numbers are not an index they are how often the key or key combinations occurs in the text/corpus. So assuming your message just now is the corpus then you used "f" 7 times (if I counted correctly" so the unigram format 'f': 8, "j": 1, and so on
Ah! Thank you for breaking that down for me! I'll ask more questions as they come up if you don't mind.
When you say German or English "text", your referencing all words in the language as the corpus?
Solid observation on neutral layout aspects btw. Man, this is a good read.not finished yet.
I'd be thrilled to see a response or any comments from Riley on this. I'm assuming you sent him a copy to check out..?
Also, does the part I highlighted in the img mean that your assuming shift is being primarily used with the right hand?
Interesting take on the dup key, as well.
Did I miss something or did you exclude your tweaked layout?
Finished it. This should be included in every cc product's box.
Thank you for your feedback! Always good to hear that the effort I put into this improved someones understanding or helped in any regard 🙂
Yes for the finger usage shift is assumed to be exclusively used on the right hand. It is just a number to get a rough idea about the distribution so no point in arguing about the correctness of it. I have included my layout in chapter 6.6.
No I have not sent it to Riley yet but let`s do it now. @tawdry sphinx I have hinted at this a while back on stream. I am curious what you think about it. And maybe even correct my assumptions about how you created the layout especially in regards to the artistic/chord optimization.
I don't personally know anyone, or noticed anyone use the right shift key, tbh. Maybe its just me but if someone asked me, I would say its pretty standard in the states for people to use the left one.
I only see one mirror key on your layout, on the left pinky, but the two mirror keys on the default layout are not identical in function. The one sentence on it (bottom of page 33) sounds like maybe you have right mirror on the left pinky and no left mirror? This could use more explanation or an example of what you are doing with the mirror key.
GG @craggy aurora, you just advanced to level 3!
I agree on defaulting to left pinky shift. My personal exception on my CC2 is I use right shift for letters on the left ring finger, which for me are I, B, and colon. I do this because I find it difficult to consistently hit only the intended character while holding only shift with the neighboring pinky.
I have never used the default layout. What is different about the default left and right mirror key? I have left mirror on the left pinky with the idea that I can make a diagonal press and use the pinky to access the layer2 and layer3 keys on the right pinky with it.
Mirror left (aka ambidextrous throwover left) essentially pulls the right hand inputs on to the left hand keys, while mirror right pulls the left hand inputs on to the right hand switches. Using your layout as an example, normally left index South will input o, but if you hold down the left mirror, it will pull the right index inputs onto the left index, so left index South will instead output d.
I only have a CC2 (my Master Forge has not shipped yet), so I don't know how feasible mirroring & selecting a mirrored input with the same finger are. It may require tweaking some tolerances, if it can work.
Ah okay I thought you mean something different. Ofc it always mirrors the other hand. I dont have a usecase for getting my keys from the left hand to my right one. I may only need the rights ones on the left hand if I use my mouse for example. Thats the only thing that came to my mind. So that I can make left hand shortcuts. Mirror on the same hand works only with a diagonal press. But that works reliable compared to the cc2.
Another reason for having both mirror keys is for chord arpeggiates. Left mirror conjugates the prior chord into present progressive form, while right mirror pluralizes. In practice, this typically means left mirror adds -ing to the end of the chord, while right mirror adds -s or -es (if the chord ended in s or a vowel).
Yes that would be a reason. But with the current implementation a useless feature for any language besides English anyway so I am just waiting for custom arpeggiation modifiers which I just throw onto a ring finger or something 🤷♂️
Do u have the 2000+ trigram timings in a spreadsheet you could share? I’m curious to see which key combos are good and which are bad
But you can click diagonally... It's just that there's a pretty small window that's hard to get in.
On M4G I think it is pretty easy. For a CC2 my layout would probably not be that effective for exactly that reason, yes.
Here are some plots of the typing data analysis. Terminology is sometimes in german but I think you won´t have a problem understanding it. All from the right hand. T=Thumb, R=Ring finger, P=Pinky, M=Middle finger, I=Index finger and then the directions.