I'll participate in an exhibition. In March 2, 2024, at 東京都立産業貿易センター台東館 (Tokyo Metropolitan Industry and Trade Center Taito Building), the name of the exibition is キーボードマーケット トーキョー（キーケット）.
The figures below are the flyers for the exibition. No English version, sorry.
Kailh Choc V1 switch has a dedicated stabilizer, but it is plate mount. Making plate and PCB both is costly and cumbersome. P2PPCB can do with it by just a frame.
How does it work? P2PPCB frame has snap jaws to hold the parts.
I wrote a lot about prototyping in P2PPCB NTCS ReadMe. So I'll leave the example of prototyping over there, and praise prototyping here.
I am old enough to have been through the dot-com bubble. I read the bubblehead writers of pulp magazines wrote about "idea". In the age "idea" looked enough to make a success. Do you remember "business method patent"? Hahaha... Yeah "business method patent" was a superior idea to make money for the patent firms! But, of course, the patent firms did their jobs to make money by the idea. Their jobs were excellent. Otherwise, who would have taken such an idea seriously?
Today we re-realize the fact that a good idea is a shadow of good jobs. Someday we may forget the fact again, but so far, we know it.
I cannot list sufficient conditions of a good job. But I know a necessary condition of it: high quality interaction.
The list above is a kind of quality. I guess that you are thinking "this is just a kind". Yes, the list above lacks the interaction with physical reality.
The list above is another kind of quality. This is also just a kind.
Make a prototype: good,
If your idea is something like "business method patent", you don't need to make a prototype. Otherwise, if it has something to do with physical reality, prototyping will give you irreplaceable quality of interaction.
Prototyping is iterative. As you iterate, you will find a better idea than the original one. A good idea is a shadow of good jobs.
As discussed in the previous posts, the best one-size-fits-all keyboard is chiclet style. For XXS for XXL size person, one-size-fits-all is not the best. Let's define such a body size for me.
Able-bodied. The hand size is normal. Not required to type super-fast. A heavy user of keyboard shortcuts. Required to type symbols much (i.e. hit number row keys and Shift keys often). No need to read the keytop legends.
The features for the body size:
Shorter key pitch makes the keyboard smaller. Hitting number row keys is also easier. Many studies suggest 17 mm pitch doesn't harm speed or error rate compared to 19 mm.
How many millimeters should the travel be? Long travel like traditional 4.0 mm is a bad habit from the non-electric typewriter era. Recent traditional looking keyboards have shorter travel than 4.0 mm. Long travel worses finger interference and typing speed. However too short travel causes poor experience. Where is the best?
Most chiclet style has 1.2 to 2.5 mm. Apple makes around 1.0 mm and I hear constant complaints for years. The best number should fall in 1.5 to 3.0 mm. A study suggests 2.0 mm. I think this is true from my experience.
R4-R3 has long step, and R3-R2 and R2-R1 has short step. R0 (nearest row) keytop angle is quite steep. Like the figures below:
I love the undulations of the step and keycaps. Chiclet style looks dull.
My Golden Fleece will have these features.
You might notice my Junana keycap cannot do 2.0 mm travel yet. AFAIK on the market there is no switch that has MX stem and 2.0 mm travel. I believe that the market of low profile switches will go to 2.0 mm travel soon. 3.2 mm is apparently too long for ergonomics yet.
In the recent posts, we saw many parameters of keyboard ergonomics.
These parameters relates each other. Some parameters are irrelevant for heavy users.
From these parameters, I reached a conclution that recent chiclet style is the end of keyboard evolution. We will never see any next style in fashion while we are alive. This is just like IBM Model M layout. If you want the best one-size-fits-all style, chiclet is it, and this will be true after a hundred years.
I want to point out the readability of legends. The age of professional typists is gone. Most people have to read legends of keytops very often. The keytop width / depth of MacBook is 16.5 mm. As mentioned in the previous post, mechanical keyboards' keycaps have much smaller keytops, and the smallness is inevitable for ergonomics.
(BTW I don't believe large keytop is good for heavy users. If so, IBM should make Model F keytops larger in 1980. Large keytop is good for readability, and this is a decisive factor.)
Of course chiclet style is good for laptops, and a human being is an animal of habituation, we are comfortable with the same style between laptop and stationary. Why do we want different thing?
The answer is simple. Because we are perverse, weird, attention seeking geeks. Comfortable? That is for couch potatoes. We can't stand one-size-fits-all keyboards and overcome any hardships to seek the Golden Fleece! We are Jason!
Calming down...
The quest of the Golden Fleece is not comfortable, but the Golden Fleece itself should be comfortable (just for the owner), hopefully more comfortable than one-size-fits-all things.
I don't expect we will reach the same conclusion about what is the Golden Fleece. If I expect so, I'll produce and sell my Golden Fleece instead of P2PPCB components. However I believe we are curious with the Golden Fleeces of others and use as references for the quest of ourselves. In the next post I will describe some (not all) features of my Golden Fleece.
In the previous post I told the importance of the keytop depth size. So let's compare the keytop depth sizes of common profiles. I made 3D models of them for P2PPCB. The figure and numbers below are the projection to the switch mounting plane.
As the theory which is described in the previous post, uniform profiles should have short keytop depth because they (usually) lack the step. XDA and DSA follows the theory.
Larger keytop size is better to make legends readable. But smaller is better for reducing finger interference (it occurs between left and right keys too). From this viewpoint, I think Cherry profile does the best. It has long depth but the step makes it harmless. Short width is good for reducing finger interference between left and right keys.
I want you to pay attention to Model M. It has smallest (depth and width both) keytop among the stepped profiles. In 1980 IBM was a giant of the industry and invested much to keyboard ergonomics. The glory of Selectric typewriter was not long ago. To be honest, I think that the decline of IBM begun with Model F, especially Olympic-podium style wide keys. However it came from cost pressure, and the knowledge of R&D was not marginal. The keytop size of Model M might be best for professional typists (in 1980 they are still primary customers). Professional typists never look at keytops, so readablity is irrelevant for them.
So my Junana profile? Shorter than Model M, 13.1 mm. The requirements of 17 mm pitch and dye sublimation (five sides at once) was strict...
Aside from XDA and DSA (uniform profiles), common keycaps are designed like the figure below (Cherry profile):
The green arrows indicate "step". This gives keyboards beautiful undulations. The origin of the step is probably non-electric typewriters like this:
The purpose of the step is to avoid finger interference while pressing:
Old non-electric typewriters have long step, and recent chiclet style keyboards have zero step. Non-electric typewriters have long travel so they require long step. Recent chiclet style keyboards are the opposite.
Keytop depth (near to far) also plays a role. You can avoid finger interference by reducing keytop size of depth direction. This is a big reason why chiclet style keyboards have gaps around each key.
Finger angle also plays a role. The step is useless if the finger is perpendicular to the switch. Therefore R4-R3 step should be larger than R3-R2. The Cherry profile figure above follows the theory. However R1-R1 step doesn't. In this case, another phenomenon is in play. See figure below:
This layout has zero R1-R1 step, and in my experience with many prototypes, this is a bad idea. I felt my fingers cramped in this layout. I felt much better below:
In near rows, the step doesn't play a role. Instead of the step, keytop angle plays a role. The keytop height should be flat in near rows. I have no idea about the theory behind the phenomenon.
BTW all the figures above are created by P2PPCB Composer F360. This is also good for making figures.
