Kinesis layout

The Kinesis keyboard is programmable. You can rearrange the key caps relatively easily with a supplied tool.

Unfortunately, there are over a dozen (subtly) different key shapes - and not every key will work at its best in every location.

The main problem is that the keys have different degrees of front-to-back slope.

Some of the key-caps are coloured. These keys are the ones that would be home keys under the QWERTY layout. You can see these keys have smaller upper surfaces - clearly an attempt to provide tactile clues to indicate the home position. After rearranging the keycaps any usefulness of these cues would be much diminished.

My layout is based on the [Dvorak] layout.

Initially, I tried the Dvorak layout - but it didn't work well for me. I didn't like the way it put the letters "A" and "S" under my smallest fingers - and had "P-U", "U-P" and "H-G" combinations - which required using the same finger twice in a row.

However - I was impressed with the Dvorak design principles - and I deliberately retained many of the features of the Dvorak layout in my own design.

For example the vowels are all in the home row under the left hand - while the five most common consonants are in the home row under the right hand.

Also I attempt to follow the "alternating-hands" principle.

It took me several days to settle on a design. I tried a number of experiments in the interim.

Retrospectively, I might swap W and K - because of WH and KN combinations.

Also U and R - since R would be nice to have on the home row, and the left hand needs to be used more.

Initially I rearranged most of the key caps. You can see photographs of the results. However this meant that many of the keys were no longer at the right slope - and that the home position cues were much diminished.

Because I felt that initial experiments with relabelling keys had been successful, I embarked upon relabelling on a larger scale - in order to return the keys to their right places. By the time I had finished, practically every key had its own label. Unfortunately, such labelling makes keeping the keyboard clean more challenging.

Feedback

There is an optional audible click on every keypress - which provides very useful feedback - and helps avoid "bottoming out".

I supplement this feedback with software on the computer that performs much the same function - but has different sounds for different keys - and has a volume control.

IBM's 'buckling spring' technology has the best feedback of any keyboard I've used. It combines both tactile and audio feedback.

Since keyswitches can include LEDs, adding illumination to this sort of keyboard is possible. If done in such a way that the LEDs could be controlled, they could be used to provide additional visual activation feedback.

Problems

• In my view, far the biggest problem is with the microswitches that lie beneath every key.

  These are much too stiff - and they bottom out too suddenly.

  "Hitting the bottom" with too much force may be a significant factor responsible for causing cumulative damage disorders of the type most customers will be seeking to avoid.

  Despite the existence of the feedback described above, I feel that the force required to activate the keys is still much too large - and it is still much too easy to bottom-out accidentally. Also, hitting the bottom of the travel is like running straight into a solid brick wall - there is no "give" at this point at all.

  I have used a number of other keyboards with better feel and consistency than this one.

• The function keys are of inferior quality. In particular, they sometimes produce their tactile click response without signalling a keypress. Also, I can't rearrange the keycaps; sticking labels to their surfaces is an ugly hack - and there's practically no space for any function key strip.

• Some way of extending the hand rest region towards the user seems to be needed. I find this region to be several centimetres too narrow.

• Remapping keys using macros has some niggles. If you remap a key this way its auto-repeat rate becomes unusably slow. I also note that pressing such a remapped key in some applications causes them to receive additional control characters - which can cause them to malfunction. This happens (e.g.) using pico via a telnet session.

  I believe this bug has something to do with the way shifted characters are replayed. It appears that a <shift-lock><key><shift-lock> sequence is generated - and the shift-lock keypress is causing problems.

  Remapping keys using macros is the only way to remap a key stroke without remapping all the other functions bound to that key.

• The firmware's API is proprietary and undocumented. There is no supplied software for many common OSs - so this is a big issue. You can't save your settings, you can't print out the existing macros - you can only change things using the supplied firmware and the manual interface. Perhaps the USB version works better in this respect.

  I have had to reprogram my keyboard from scratch many times. Here's what I have to do each time. An interface for doing this programmatically seems highly desirable to me.

• Using 48 macros does not seem to work properly. Kinesis acknowledge that there are some minor issues - but I encountered major bugs that made employing the 48 macro mode into an unthinkable proposition.

  The problem was lost keystrokes. Not just when hitting two keys in quick sucession (as [the support site suggests]) - but all the time. I could easily press a key 30 times and still not have it register once. Shift and control key operations only exhibited the problem.

  Fortunately, reverting to 36 macros cures the problem.

• Bugs in the firmware mean that the shift keys sometimes act as thopugh they are stuck down. This is pretty irritating - until you find how to work around the problem.

Wishlist

• More keys. 86 keys isn't anything like enough. It's great being able to program keys - but a pain having to bind them to key combinations for lack of physical keys.

  I think that an improvement would be to remove the existing function keys - and to add on two additional rows of keys, directly above the existing ones.

  There also seems to be some space left for keys in the bottom corners of the keyboard.

• Also, I would like more thumb keys. I'm convinced that space could be found for these. It seems to be important to have your main modifier keys thumb-operated. As it is I have sacrificed the functionality of the two "Alt" keys - and turned them into additional "Shift" keys.

• After years of searching I finally found some keycaps suppliers.

  I originally thought that ripping to bits a Cherry G80 series keyboard and scavenging its keycaps might work - but those keys are all different shapes as well: this is not practical.

  The POS-style technology (as used by [X-Keys] looks good to me.

  The [X-Keys] keycaps are interchangable with the Kinesis ones - because they use exactly the same microswitches. Unfortunately, 'International Shipping' on a handful of X-Keys keycaps comes in at $49.49.

  After quite a search I found a UK supplier: [RS components].

  Photos of my keycaps can be found here.

Keyswitches

Alas, spring surgery on the fingerboard keyswitches looks as though it will not be practical to perform.

This problem is a serious one for me - and it probably means that my Kinesis keyboard will eventually be retired.

Delivery

The keyboard took four weeks to arrive by surface mail.

Latest keymap

Links

This site

Relating to this page

Remapping instructions for the current layout are available.

Kinesis