AcidFire's modular keyboard system - Axios [In Development]

[Koren]

I blame schools for teaching kids to use computers but not teaching them how to type.

I've been taught the basics of touch-typing when I was 10 at school. Using a neat program where a conveyor belt was bringing letters, and you had a trap in the middle that could catch the letter should you press the right key at the right moment. First, it was the keys on the home row, then you were working with the letters all over the board. It's something like 26 years in the past, but I can still see the yellow letters flowing on the blue background. ^_^

I'm pretty sure that many kids of my age at this time were given the opportunity to do the same here.

That being said, I don't think that touch typing is a more valuable lesson than most of what's done in schools.

I'm not even convinced that touch-typing is so important for most peoples (although I find it obviously valuable myself).

With the traditional QWERTY keyboard I had to read a typing course book to figure out which fingers should press which keys.

There's different views about which fingers should be used for each key, and I'm not even convinced that there's a definitive finger that should be use on a key (like some really proficient typist, by the way). I created my "own" way to touch-type, although I read later my mother's books out of curiosity (that differ a bit from the usual suggestion, but I like them better because they are closer to symmetry, using for example middle finger for X).

I'd say I can live with the ISO-ANSI being a standard for home computers, since I can still choose the variation I like the most. The issue is indeed with laptops, and I don't believe I'll see the day where we'll see a decent keyboard on a laptop, unfortunately.

[clickclack123]

I agree with these two. I have both an ergodox and a truly ergonomic.

Did you buy the ErgoDox after the TEK because you were looking for something better? How do the two compare (pros cons about both)?

I had the Ergodox first. I bought the TEK because I wanted something to use at work that was similar to the Ergodox, but slightly less "out there" so I wouldn't get so many questions from people. Also I kind of don't trust some of the people who have access to my office after-hours, so I didn't want to leave something that I've put so much effort into making like the Ergodox around... I can replace a TEK but the Ergodox would be a pain in the ass to replace.

I do like both of them but in different ways.

Ergodox Pros:

• Massive, ridiculous programmability and customizability, especially using TMK firmware. Even crazy stuff like mousekeys, etc.
• Many useful thumbkeys/lower row keys/center column keys on both hands. Basically all the keys that I could want, and more than I can remember 
• Excellently suited for stenography using Plover
• The satisfaction of using something you've built and programmed yourself
• Choose your own keycaps and switches (I really like the feel of the DCS keycaps that I got from Massdrop, much better than the DSA that I started with)
• Matrix, non-staggered keys. I will not go back to a wonky staggered keyboard ever again.

Ergodox Cons:

• No portability at all with separate parts and trrs cable, also seems more fragile with the trrs connectors
• (might not be a problem for other people) thumbkeys hanging lower makes it a bit awkward for me to use my Rollermouse Free2 below it

TEK Pros:

• One-piece, more "standard" looking design
• Matrix, non-staggered keys. I will not go back to a wonky staggered keyboard ever again.
• Nice solid-feeling case

TEK Cons:

• Very questionable reliability, no support at all from TE, and my new TEK has had a lot of problems with double-pressing keys/missed keypresses, that I've detailed in another thread. It's not so bad now it's "broken in" and I've mucked around with firmware latency settings, but it shouldn't have problems at all IMO, and it's the only keyboard I've ever known to need a "breaking in" period for MX Brown switches to work. The Ergodox def feels more "snappy" in my case (5ms latency on the TEK, I think the Ergodox is 2ms? Whatever the default for TMK firmware is). My Ergodox or any of my other mechanical boards have never ever had a missed press or double press.
• Limited customizability (you can make your own layout, but can only have one function layer). I would like it if I could turn the End key into a different function layer, maybe to make the right hand into a NumPad.
• Keycaps feel a little "cheap" to me, and a bit heavy, especially the large keys such as Enter and the center column keys (think they're ABS). Both my boards are MX Browns but the Ergodox keys feel a lot lighter, and they're actually significantly quieter for some reason as well. Could be because of the plate in the TEK? My Ergodox has no plate. The Enter key on the TEK is pretty loud IMO.
• Weirdly shaped, un-replaceable larger keycaps (such as the Enter key, center column keys, and Spacebar)

If I had to choose only one, it'd be the Ergodox for sure. The TE is a compromise that I got because it's a bit more suitable for my work environment.

I don't actually use the larger keys in the middle of either the Ergodox or the TEK much. I think that's partially because I'm constantly switching between the two boards, so any keys on the layouts that are different between them, I tend not to use or remember what they're programmed to.

I like to hit the left- and right-most keys (shift keys on a normal keyboard) with my palms, at the base of my pinkies, so I normally would map those to be the Enter key. Looking at the US version of the TEK, and also AcidFire's keyboard, these keys are slightly wider which means they may be easier to press this way by a range of users with varying hand sizes.

For me the centre column has always been (from top to bottom) Win key, Ctrl, Alt. On the ErgoDox there are two larger keys straddling the home row vertically, which is not the case on the TEK or AcidFire's keyboard, I wonder which one feels better in practice.

I personally prefer the Ergodox layout for the center keys but that may be because that's what I started with. I find it more natural to just reach over with my pointing finger on both hands, feels like a bigger target. Maybe you could get used to either though?

Show Image

Show Image

Also, could you comment on the enclosure? The ErgoDox one seems kind of boxy with those sharp corners, while the TEK one is curvier and somehow more "inviting", so to speak.

I thought the same thing about the case being boxy, so I bought the czarek case for my Ergodox that seems much more compact, so I can't comment on the standard Ergodox case because I've never had one. I'm very happy with the czarek case. I do like the solidity and feel of the rounded edges of the TEK case. The whole thing feels nice and solid. Unfortunately for me, my TEK wasn't made as well electronically as it was physically.

I have programmed the left arrow key on the teck so it is a Fn key that I hold down with my thumb to turn the home row into arrows. So I don't even have to move my hand, I just press the thumbkey and J is left, K is down, L is right, and I is up. Also H is Home, ; is End, U is PgUp and O is PgDn.

My ergodox is set the same, using one of the thumbkeys.

I use dvorak so it's not actually those keys, but you get the idea.

I could not go back to a layout without such a setup anymore. It's so useful being able to navigate without moving your hand.

That's really interesting. I did a prototype layout yesterday on a Japanese HHK2 Lite which has a bunch of extra thumb keys. I moved all keys of the right hand to the right by one key, so I get an extra column in the centre which now contains Ctrl, Alt and Win (so it is basically a poor-man's TEK).

When I press and hold the left thumb modifier key I get numbers under my left hand and arrow keys under my right. I am finding key combinations such as Shift+Home (which on a normal keyboard only involve two keys) hard to get the hang of, because of the extra modifier key involved. I think my problem is that the way I've set it up I have to press Shift and the modifier with the same hand, the way you've set yours up you are probably pressing the modifier with the right thumb and Shift with the left hand, am I right?

That's correct. I still find it a bit difficult when I want to select the last word (shift-ctrl-left) and things like that though. My brain finds it harder when I have to do two modifiers. But any combination using a single or no modifier feels natural to me.

BTW the thumbkey also turns the numbers into Fkeys, which I also find very natural because the Ergodox doesn't have Fkeys, and I know how to touchtype numbers.

I've been using layers for symbols and numbers for many  years (since 2002 actually), but I always had my arrow keys mapped to the numbers row, so maybe I have grown used to the immediacy of dedicated arrow keys vs using a layer for that. Therefore I'd like to hear about your adaptation experience.

I found it super easy once I set up that thumb navigation layer. Never had any trouble getting used to it.

[AcidFire]

Wow, lots of awesome discussion going back and forth with this! I've been reading and trying a few things quietly, and I think for the initial layout we'll be going with a bit more of a normal bottom row. For those of you wanting a more traditional arrow cluster, it's marked down as the first layout option to be released after the initial product run is done. Also thank you Clickclack for the break down between the ergodox & the TEK, immensely helpful.

For those of you who don't make use of twitter, I've recently been accepted as part of the pre production testing for a new bit of cloud software called Onshape. Onshape is basically like Solidworks but delivered via the browser instead of a desktop app, with all the rendering done on their servers. I've been assured that there will always be a free version (fully featured too), which means those of you looking to build or modify your own designs will have an easy & free tool to do so (including the ability to fork any design I've shared). Unfortunately an NDA prevents me from saying too much about the software, I will tell you that it's been pretty fantastic to build in and I'm hoping this is the defacto tool I use for Axios right up to production.

Also on the design side, I wanted to share a little bit of my thought processes and why the casing seems to be taking so long to finish. I've said from the start that it wouldn't go to crowdfunding until I was 80-90% happy with the design; this used to be "production ready", but I've since learned that especially with user feedback that won't happen right out of the gate. I do however want to keep the dev cycle after funding to an absolute minimum, hence the seeming "perfectionism".

Up until I switched over to an adjustable design, things were relatively straight forward; this changed quite a bit however when I did. In order to accommodate the adjustable design, I lost roughly 30-40% of the space you would normally find in a standard mechanical keyboard case. This was to allow the arms to move in and out as well as up and down, all while using standard hardware to do so.

 Connectors have been the other major design consideration. I've had input devices that come with custom cables and once one of those disappears in a move or even dragging it to and from school/work/lan parties/wherever, they're either impossible to replace or ridiculously expensive for what they are. While the original plan was to go with TRRS cables like the ergodox, I've seen enough complaints about failed connectors/cables that I started looking at other options. I've since landed on using phone handset cables, often mislabeled as RJ9/10/11. They're easier to find than TRRS cables, more robust and have a greater number of options for things like curled for less clutter. They also make a more solid connection and more importantly, there's no chance of shorting something out while you're plugging them in (a risk with TRRS). There's only one problem with them: the size of the modular jack.

Modifying the case design to fit them meant I had to add another 3mm to the height of the case, which is never ideal. There are a few positives though; I can now use standard 3mm thumbscrews, and it gave me enough room in the case that the folding stand can now sit flush with the bottom of the case when not in use. A bonus on top of this is that the 1/4"-20 mount is now hidden by the stand when packed up, protecting it for travel. Also worth mentioning is the additional space for these connectors also opens the doors for others, like the potential ability to use ethernet cables for bridging or for the use of a full size B connector (seriously, I've had more than a few people ask).

The other problem with using the modular jacks and their height is their length. While only 6mm longer than the TRRS connectors, in design this small that's a massive difference especially since their height would mean having to increase the thickness of the case further to prevent potential shorts between the legs of the switches and the PCBs for the connectors. To counteract that, I've made the difficult decision to eliminate the f-row mounts and increase the space at the top of the unit by roughly the amount it would have taken up. It's not all bad news however; not only does this make enough room for the connectors, I have enough space left over to add the missing 3 keys from the thumb cluster to the empty space not set aside for the connectors. It also allows for a larger, more stable folding stand and room for a bigger battery for the wireless option.

As for the now orphaned f-row keys, the increase in the depth would allow for another module the size of the thumb cluster to be mounted there, and with allowing for some space for the RGB controller, you would be able to add an additional 6-9 keys

On top of all this, I'm still trying to balance the aesthetics, because while I know we all value function (and RSI relief) over looks having a nice looking board helps to get the public on board. During the events I attended last year with prototypes, I got more positive "must have" feedback from the general public than I ever would have expected. And this is absolutely the sort of thing we want to see, the more boards we sell, the better rates I can get for parts & manufacturing and thus the cheaper I can sell them. And more importantly, just that many more people we're helping and educating that the current standards suck and there are better options out there.

Hopefully this is all clear, and if it's not don't fret as I'll have additional photos soon (hopefully renders as well when if i can get permission to post screen caps from Onshape).

[jeep]

 Thanks for the update!

[JackMills]

Interesting insight in the design decisions, thank you for sharing AcidFire.

[GenuineMP5]

Just found out about this today. Spent the last couple hours reading through the entire thread. While this version would have been the one I wanted, I certainly see how something like the current version would appeal to a larger crowd.

AcidFire, absolutely amazing work all around! Now, I'm going to go play with Lego and pretend I'm anywhere near as skilled as you.

Looking forward to watching progress of this!

[solarundies]

It's awesome to hear that you're back working on this! I'm excited to see your new design choices.

[AcidFire]

In the last week or so, I've had a couple of epiphanies (hurray for long drives home!) that have not only cleared up some electrical design issues, but more importantly reduced the cost to produce the casing by a substantial amount.

First, for those of you who were missing the three keys on the thumb cluster they've been added back, in a bit of a different spot:


And as I mentioned previously, the f-row/hotkeys now have a new home:


With the plan to release something down the road to connect the halves in the center, we'll end up with a layout that looks something like this:


As I mentioned, I believe I've found a way to eliminate the need for slides from the injection molds, which is a not insubstantial savings -



I've also been working on & testing thumbscrews -

Both of these have been very promising

Along with the case design I've been finalizing the connectors. The kits will most likely come with Mini/Micro USB, 6p4c connectors and a blank or two. However for the beta unit we'll be including all of them for extensive feedback from you guys.

Top Row, left to right:
- 4p4c (Phone Handset Cables RJ11Mini)
- 6p6c (Phone Wall Cable RJ11)
- 8p8c (Ethernet Cable)
Middle Row, left to right:
- 3.5mm TRRS Cable
- Mini/Micro B USB
- Full Size B USB
Bottom Row:
- Blank Cap

I'll post more as they come off the printer.

[Faokryn]

Well, you can count me among those who registered an account to comment on this amazing project.  :]

AcidFire, first of all, your commitment to open source and to accommodating those with disabilities is commendable and inspiring.  Not to mention how inspirational this project as a whole is: reading through the thread made me want to get out and make things!

I found a link to this thread on a blog post regarding the ErgoDox, and referencing the original ErgoGP design.  Since learning about the ErgoDox a few months ago, I've been excited about potential modding capability, but I think the Axios's modular design will be even better, allowing the creation of custom modules (and releasing them open source for anyone who wants, of course :] ).  Now I'm watching this thread, subscribed to the crowd supply, following on Twitter, and among those eagerly awaiting each update!

I do have a couple of questions / comments regarding some of the most recent design changes.

3) Thumb cluster switch changes
The more I look at the design, the top three 1u switches on each side seem to be counter intuitive to the end goal of reducing strain by reducing travel, so I'm experimenting with dropping them. In their place, I'm looking at using the space for a set of 5 RGB LEDs on each side for indicators, since dropping the switches & caps helps to balance out the cost while putting them in a nice visible position. For those of you asking for some kind of display, this would also make for a very handy mounting point.

I have enough space left over to add the missing 3 keys from the thumb cluster to the empty space not set aside for the connectors.

So one row of 1u keys has been moved from the thumb cluster to the top of the keyboard, and replaced by indicator LEDs?  It seems like the keys would also be difficult to reach at the top of the keyboard.  Isn't it better to have the extra keys on the thumb cluster, even if they're secondary to the other two rows? Perhaps moving the LEDs to the top of the keyboard instead?  I was under the impression that having more thumb cluster keys is optimal, because it allows you to have more keys for changing layers while keeping the rest of your fingers on the home keys.  I'm interested to hear some opinions regarding that from people who have ergonomic keyboards (I don't).

Anyway, I'm sure we'll see the all-key thumb cluster module down the road, if not as a stretch goal for the crowd funding campaign, then soon thereafter.  I for one wouldn't mind having one all-key thumb cluster and one indicator LED one (the opinion has already been expressed that 5 LEDs is more than enough for many people).  That's the best part about this system!

[JackMills]

I am curious to see how the new hotkeys cluster looks like when tented. Good job on improving your design for manufacturing, most people don't realize how important this is, so kudos on that. I always liked this total approach of you and I am very anxious to get hold on one of these, but I am also willing to wait because I think you could produce something very nice.

[Koren]

AcidFire, I may have thought one second about cursing you for making me adapt again my layout, but...

But it's just great, as usual. Can't wait, really.

Using the available space on top to put back the buttons from the thumb cluster is nice (it's maybe even better for me, in fact). I'm curious to see how the "f-keys" add-on can be secured in place when folded strongly enough, but I'm sure it'll be nice.

Thanks for all your hard work, and I hope that the accident is just a bad memory for your family now.

Isn't it better to have the extra keys on the thumb cluster, even if they're secondary to the other two rows? Perhaps moving the LEDs to the top of the keyboard instead?  I was under the impression that having more thumb cluster keys is optimal, because it allows you to have more keys for changing layers while keeping the rest of your fingers on the home keys.

The third row on the thumb clusters was quite difficult to reach, anyway. You probably can't use them during "normal" typing. I think "out of reach" keys still have their use (for a pause key, a behavior-tweaking key, etc.), so that's great we don't lose those, I think.

As far as layers are concerned, if you combine both sides, you could already have 49 (7*7) layers directly available without even pressing two keys with a thumb (which is definitively possible for adjacent keys). And that also don't take into account that a single key can be used at the same time as a layer switch and a dead key, so you can have several hundreds layers available easily. Probably more than you can remember...

[AcidFire]

Well, you can count me among those who registered an account to comment on this amazing project.  :]

AcidFire, first of all, your commitment to open source and to accommodating those with disabilities is commendable and inspiring.  Not to mention how inspirational this project as a whole is: reading through the thread made me want to get out and make things!

I found a link to this thread on a blog post regarding the ErgoDox, and referencing the original ErgoGP design.  Since learning about the ErgoDox a few months ago, I've been excited about potential modding capability, but I think the Axios's modular design will be even better, allowing the creation of custom modules (and releasing them open source for anyone who wants, of course :] ).  Now I'm watching this thread, subscribed to the crowd supply, following on Twitter, and among those eagerly awaiting each update!

Awesome, glad to have you aboard! If you have it, would you mind sharing the blog post where you found the link?

As for your questions...

So one row of 1u keys has been moved from the thumb cluster to the top of the keyboard, and replaced by indicator LEDs?  It seems like the keys would also be difficult to reach at the top of the keyboard.  Isn't it better to have the extra keys on the thumb cluster, even if they're secondary to the other two rows? Perhaps moving the LEDs to the top of the keyboard instead?  I was under the impression that having more thumb cluster keys is optimal, because it allows you to have more keys for changing layers while keeping the rest of your fingers on the home keys.  I'm interested to hear some opinions regarding that from people who have ergonomic keyboards (I don't).

Anyway, I'm sure we'll see the all-key thumb cluster module down the road, if not as a stretch goal for the crowd funding campaign, then soon thereafter.  I for one wouldn't mind having one all-key thumb cluster and one indicator LED one (the opinion has already been expressed that 5 LEDs is more than enough for many people).  That's the best part about this system!

I probably should have been more clear about the strain I mentioned regarding removing the top row of thumb keys. While they're not completely terrible when the board lays flat, when tented and dropped they introduce a fairly severe ulnar deviation and given that the core of this project is ergonomics above all else it goes directly against the device's goals. Even as a key you'd occasionally strike by lifting your hands, you're now introducing Pronation strain as you rotate your wrist around to hit it. With the keys in their new home, it's a shoulder-elbow-wrist movement similar to lifting and striking the F row, or a mouse movement with a substantial reduction on strain.

For those who absolutely must have three rows of keys, the 9x9 hot key cluster can be used in the thumb cluster as well, with the trade off being a reduced hit box for each of the keys. As for a module down the road adding them back, I honestly don't see that as something we'll release since a change in the case design to strengthen the cap that surrounds the keys on the top (and prevents light leaks) has shortened the 45° angled side, requiring a reduction in the length of the thumb module. I could leave it at the original length, however that adds to the cost of the device since every part of the thumb cluster is designed to be used for both halves of the board. While adding another cavity to a mold isn't a huge expense, there's also the requirement to now stock both and unlike the main case which are very clear in their difference, the thumb bottoms are less so and I would hate to accidentally ship the wrong part. By pure serendipity, the vertical side where the main clusters can connect was lengthened to nearly match the 45° side, negating the need for an extra set of molds for the hot keys.

Regarding the indicator LEDs, you may notice that I've reduced them to three per side. When I had my little epiphany (or realization that I was being stubborn) and figured out that I don't need LED control electronics in the thumb clusters, I also discovered that with the 2x3 1.5 layout that meant that I would have three RGB LEDs free. So, cost savings (component count reduction) + easier programming = win. As part of the development with the backlight, I've decided that you'll be able to assign three different behaviors to each key's backlight individually.
- Open - Answers to both backlighting/animations & indicator calls, with indicators taking priority by default (possible to make it user selectable to prioritize backlight as well)
- Backlight - Answers only to backlighting calls.
- Indicator - Answers only to indicator calls. These will be user assignable as well, with certain functions built into the keyboard and the desktop software augmenting it (for things like tweets or email notifications).

I am curious to see how the new hotkeys cluster looks like when tented. Good job on improving your design for manufacturing, most people don't realize how important this is, so kudos on that. I always liked this total approach of you and I am very anxious to get hold on one of these, but I am also willing to wait because I think you could produce something very nice.

I am as well, I'm still learning the assembly side of Onshape so modeling the tented outcome has not been happening for me. That being said, I will admit that I'm not the biggest fan of the way it all looks flat, with the exception of the two sides connected. Hopefully I should have all my parts printed by the end of the weekend and we'll see it in the flesh.

And I very much appreciate the support and patience while I dial these things in, while I'm also very anxious to get it into everyone's hands I also don't want that to happen until I know for sure that the end product is feasible. I've watched far too many crowd funded (kickstarters especially) drag on for months or years because they didn't take this time to do it right, and I am adamant that won't happen with Axios.

AcidFire, I may have thought one second about cursing you for making me adapt again my layout, but...

But it's just great, as usual. Can't wait, really.

Using the available space on top to put back the buttons from the thumb cluster is nice (it's maybe even better for me, in fact). I'm curious to see how the "f-keys" add-on can be secured in place when folded strongly enough, but I'm sure it'll be nice.

Thanks for all your hard work, and I hope that the accident is just a bad memory for your family now.

Ha I wouldn't blame you, I've mentally smacked myself a couple of times each time I change it because it means I also have to update the PCBs as well. I've learned to wait and handwire until I'm satisfied with the end result.

I'm not concerned about the strength of the mount for the hot keys cluster as the new thumbscrews make it very easy to tighten down, and short of HULK SMASH style typing they won't move (though it is designed to give with enough force to prevent the mounts from being damaged).

Isn't it better to have the extra keys on the thumb cluster, even if they're secondary to the other two rows? Perhaps moving the LEDs to the top of the keyboard instead?  I was under the impression that having more thumb cluster keys is optimal, because it allows you to have more keys for changing layers while keeping the rest of your fingers on the home keys.

The third row on the thumb clusters was quite difficult to reach, anyway. You probably can't use them during "normal" typing. I think "out of reach" keys still have their use (for a pause key, a behavior-tweaking key, etc.), so that's great we don't lose those, I think.

As far as layers are concerned, if you combine both sides, you could already have 49 (7*7) layers directly available without even pressing two keys with a thumb (which is definitively possible for adjacent keys). And that also don't take into account that a single key can be used at the same time as a layer switch and a dead key, so you can have several hundreds layers available easily. Probably more than you can remember...
[/quote]
Pretty much this. I think your title going forward Koren will have to be "Keeper Of Layouts"

[Faokryn]

If you have it, would you mind sharing the blog post where you found the link?

Absolutely! http://adereth.github.io/blog/2014/02/12/building-an-ergodox/

The third row on the thumb clusters was quite difficult to reach, anyway. You probably can't use them during "normal" typing. I think "out of reach" keys still have their use (for a pause key, a behavior-tweaking key, etc.), so that's great we don't lose those, I think.

I probably should have been more clear about the strain I mentioned regarding removing the top row of thumb keys...

Well, I trust your expertise, and your experiences using the prototypes!  Thank you for the further information.

And Koren, I agree that "out of reach" keys have their uses 

For those who absolutely must have three rows of keys, the 9x9 hot key cluster can be used in the thumb cluster as well, with the trade off being a reduced hit box for each of the keys. As for a module down the road adding them back, I honestly don't see that as something we'll release since a change in the case design to strengthen the cap that surrounds the keys on the top (and prevents light leaks) has shortened the 45° angled side, requiring a reduction in the length of the thumb module. I could leave it at the original length, however that adds to the cost of the device since every part of the thumb cluster is designed to be used for both halves of the board. While adding another cavity to a mold isn't a huge expense, there's also the requirement to now stock both and unlike the main case which are very clear in their difference, the thumb bottoms are less so and I would hate to accidentally ship the wrong part. By pure serendipity, the vertical side where the main clusters can connect was lengthened to nearly match the 45° side, negating the need for an extra set of molds for the hot keys.

Using the 9x9 cluster would meet what I was thinking of just fine!  Though with your and Koren's advice, I realize it might not be comfortable.  Still, my curious nature means it's probably something I'll have to try out, as long as I'm correct in thinking the normal thumb cluster and the hot key cluster will be easy to swap, since they'll be using the same mold 

Regarding the indicator LEDs, you may notice that I've reduced them to three per side. When I had my little epiphany (or realization that I was being stubborn) and figured out that I don't need LED control electronics in the thumb clusters, I also discovered that with the 2x3 1.5 layout that meant that I would have three RGB LEDs free. So, cost savings (component count reduction) + easier programming = win. As part of the development with the backlight, I've decided that you'll be able to assign three different behaviors to each key's backlight individually.
- Open - Answers to both backlighting/animations & indicator calls, with indicators taking priority by default (possible to make it user selectable to prioritize backlight as well)
- Backlight - Answers only to backlighting calls.
- Indicator - Answers only to indicator calls. These will be user assignable as well, with certain functions built into the keyboard and the desktop software augmenting it (for things like tweets or email notifications).

Cool modification! I love that you're constantly squeezing more efficiency out of your design!  And even with three lights, that's 8 indication states without even changing the color.  That's quite versatile!

As far as layers are concerned, if you combine both sides, you could already have 49 (7*7) layers directly available without even pressing two keys with a thumb (which is definitively possible for adjacent keys). And that also don't take into account that a single key can be used at the same time as a layer switch and a dead key, so you can have several hundreds layers available easily. Probably more than you can remember...

I meant more as single-key modifiers to use quickly while typing, but I understand that you have practically limitless layer mapping when combining keys.  I was even thinking of maybe having a "layer layer" (activated by one of the "out of reach" keys) that would map a set of keys to then switch to other layouts for specific applications.  That could be nifty! 

[Koren]

I meant more as single-key modifiers to use quickly while typing

I don't think you should be worried...

If I understand you well, that's still 12 alternate layers directly availables with a single thumb (without pressing "between" keys) and 36 by using both...

Following a Zipf law, each layer tend to be really roughly ~50 times less useful than the previous one in a given context (depending on how keys you actually use on the main part of the keypad), so even if you'll probably need the first three layers on a regular basis, the fourth one should be useful each hour, the next one each week, then each year, a couple a time in a life, then never. Three layer switchs on both thumbs are probably sufficient to store your needs for a lifetime.

Though there's a catch. "In a givent context" is important : you could probably want different layers when you're changing the language (russian, asian languages...), the application (for shortcuts), etc. For example, I'd like a key giving me α in a "normal" context but "\alpha" in a "LaTeX" mode (having a greek-LaTeX macro layer will probably spare me hours each year). But hitting an additionnal key to switch the context, or a harder to reach one, is not a problem.

I'd suggest you to try designing your layout, you'll probably run out of ideas before running out of layer switchers ^_^

I was even thinking of maybe having a "layer layer" (activated by one of the "out of reach" keys) that would map a set of keys to then switch to other layouts for specific applications.  That could be nifty! 

Actually, that's totally what I intend to do ^_^

[Faokryn]

I don't think you should be worried...

If I understand you well, that's still 12 alternate layers directly availables with a single thumb (without pressing "between" keys) and 36 by using both...

Following a Zipf law, each layer tend to be really roughly ~50 times less useful than the previous one in a given context (depending on how keys you actually use on the main part of the keypad), so even if you'll probably need the first three layers on a regular basis, the fourth one should be useful each hour, the next one each week, then each year, a couple a time in a life, then never. Three layer switchs on both thumbs are probably sufficient to store your needs for a lifetime.

Oh, I'm not worried, personally.  I have never had a keyboard with thumb switches before, but I have some friends who do (they're the one's who pointed me towards the ErgoDox in the first place, which led me here) and I had gotten the "more thumb switches == better" impression from them.  I don't really know exactly how I'll use it yet, and as you have said, just a few is probably more than I'll want or need.  I was more just curious to hear people's thoughts, given my previous impression.

Honestly, I'm much more likely to use the layers to organize and use macros than I am to have a bunch of modifiers to use while typing.  <anecdote> A few years ago, I was really heavily using the website tumblr, and I was very meticulously tagging my posts.  I had an old USB numpad, and I used the software Controller Mate to make a bunch of macros to help me tag things more quickly.  Most of the buttons on the numpad represented a show, comic, or video game about which I often reblogged art or other content.  Pressing one of these keys printed the name of the series and remapped the numpad so the keys represented characters from the series, or, if there were a lot of characters, groups of characters (antagonists and protagonists, factions, whatever made sense for the series).  I also had some keys that would print the names of every character in a group or of a bunch of characters that often appeared together in posts.  And of course I had two keys which always represented "go back one step" and "go back to the beginning" respectively.  It was pretty useful given how much I used the site back then. </anecdote> I don't plan on setting that up again, but it shows you the kind of stuff for which I have a tendency. 

I'd suggest you to try designing your layout, you'll probably run out of ideas before running out of layer switchers ^_^

I really should.  But first I have to decide if I'm going to stick with QWERTY, or if I'm going to use the switch to an ergonomic keyboard as an opportunity to switch to an alternative layout, like Colemak or Programmer Dvorak. 

[Niomosy]

I blame schools for teaching kids to use computers but not teaching them how to type.

For my oldest, typing is a requirement for the elementary school and something they do as homework on a regular basis.   That's a good thing as we were encouraged to type but there was never a really strong emphasis put on it in any computer class I've taken in elementary school or Jr. High back in the days of the Apple IIe in schools.

[Scoox]

I blame schools for teaching kids to use computers but not teaching them how to type.

For my oldest, typing is a requirement for the elementary school and something they do as homework on a regular basis.   That's a good thing as we were encouraged to type but there was never a really strong emphasis put on it in any computer class I've taken in elementary school or Jr. High back in the days of the Apple IIe in schools.

That's nice to hear. A lot of schools don't teach practical skills.

[pdf]

I've made the difficult decision to eliminate the f-row mounts and increase the space at the top of the unit by roughly the amount it would have taken up. It's not all bad news however; not only does this make enough room for the connectors, I have enough space left over to add the missing 3 keys from the thumb cluster to the empty space not set aside for the connectors. It also allows for a larger, more stable folding stand and room for a bigger battery for the wireless option.

As for the now orphaned f-row keys, the increase in the depth would allow for another module the size of the thumb cluster to be mounted there, and with allowing for some space for the RGB controller, you would be able to add an additional 6-9 keys

I've been really looking forward to this keyboard, however I was already a little worried about how easy to reach the function keys were going to be, now with them moved to be largely inaccessible, I think this is a deal-breaker for me - they're an integral and frequent part of my daily workflow 

[QuadGMoto]

I think I need to lobby for TECK style diamond key arrangements for the arrow, page up/down, and home/end keys. I know the thinking is that making them linear in the bottom row makes it possible to hit them without moving your hands, and that may be true.


But how many times do you hit those particular keys just once? How many times do you hit only one of those keys? Is it really feasible to reach two rows down without moving your hands? It seems to me (IMNSHO) that even with "in line" arrangements of these keys, moving your hands to reach them is still normal. Heck, even in my current TECK, the keys are in range for normal typing, but moving to reach them is normal because of how they're used.


Because of how they're used, they're also highly directional. But laying out such directional keys in a line means you either have to memorize something that's counter-intuitive (move left to go up, right to go down, or visa-versa) or look back and forth between the keys and screen.


Bottom line, if you're going to be moving your hands to use those keys anyway (actually more of a shoulder movement than hand movement), then the keys should be laid out in the most intuitive and quick to use arrangement.


I'm still excited about this project, and this probably wouldn't be a showstopper for me. But I do feel quite strongly about this. 


I haven't seen any discussion of a paired numeric keypad for a while. (Really important when typing lots of numbers, and I do.) Is that going to be part of this project?

[hoggy]

I understand your point about the arrangement of navigation keys, but once I got used to those keys on my kinesis advantage... :-)

[sordna]

I understand your point about the arrangement of navigation keys, but once I got used to those keys on my kinesis advantage... :-)

Same here. I am fully accustomed to regular arrow clusters and Kinesis Advantage arrow keys, and honestly I prefer the Kinesis Advantage arrangement.

[jleechpe]

I think I need to lobby for TECK style diamond key arrangements for the arrow, page up/down, and home/end keys. I know the thinking is that making them linear in the bottom row makes it possible to hit them without moving your hands, and that may be true.


But how many times do you hit those particular keys just once? How many times do you hit only one of those keys? Is it really feasible to reach two rows down without moving your hands? It seems to me (IMNSHO) that even with "in line" arrangements of these keys, moving your hands to reach them is still normal. Heck, even in my current TECK, the keys are in range for normal typing, but moving to reach them is normal because of how they're used.


Because of how they're used, they're also highly directional. But laying out such directional keys in a line means you either have to memorize something that's counter-intuitive (move left to go up, right to go down, or visa-versa) or look back and forth between the keys and screen.


Bottom line, if you're going to be moving your hands to use those keys anyway (actually more of a shoulder movement than hand movement), then the keys should be laid out in the most intuitive and quick to use arrangement.

On the one hand I agree.  When at work with a generic HP keyboard I am very used to just dropping my hand over to the arrow cluster (inverted T) and using it to navigate.
On the other hand, setting it up like it is on the Kinesis is actually quite easy to get used to.  One hand is left-right, the other is up down.  Other than it being two handed, it is essentially the same as the HJKL cluster in VI for moving around (once your fingers know it there isn't any extra effort involved).

I haven't seen any discussion of a paired numeric keypad for a while. (Really important when typing lots of numbers, and I do.) Is that going to be part of this project?

I suspect the solution to this will be similar to the Kinesis situation.  Some sort of Layer Toggle that changes the right hand cluster into a 10-key (thumb keys for 0 and Enter, various mathematical options on number keys and to the right of the 'numbers'.

[Faokryn]

I think I need to lobby for TECK style diamond key arrangements for the arrow, page up/down, and home/end keys...

...Because of how they're used, they're also highly directional...

...the keys should be laid out in the most intuitive and quick to use arrangement...

I, for one, probably won't even use those keys for arrows, page up/down, and home/end, opting instead to use a layer to map the vim directional hjkl to arrows etc.  Placing the keys in a cross has an implicit directional connotation, which, as you mentioned, makes sense if you're going to use them for directionally-associated functions, but it's a very counter-intuitive layout for anyone planning on using them for anything else.  With the newer configuration, the keys don't have that implicit directional connotation for anyone planning on using them for other things, but (unless I'm mistaken) will still be in a somewhat directional layout, similar to the inverted T, for anyone who does want to use them for those functions.

AcidFire, I think you're going to find this one a very divided issue with strong opinions on both sides.  I'd advise you to go with the configuration that you want, and the rest will have to wait for a swap-in board down the road.

[AcidFire]

I've been really looking forward to this keyboard, however I was already a little worried about how easy to reach the function keys were going to be, now with them moved to be largely inaccessible, I think this is a deal-breaker for me - they're an integral and frequent part of my daily workflow 

Unfortunately I think this is the case with any staggered design that doesn't move the f keys to match that pattern (of which I don't actually think I've seen any). I hope you'll still be willing to try a unit when they're done, maybe the new cluster or even a remap will end up being to your satisfaction.

I think I need to lobby for TECK style diamond key arrangements for the arrow, page up/down, and home/end keys. I know the thinking is that making them linear in the bottom row makes it possible to hit them without moving your hands, and that may be true.
- snip -

I, for one, probably won't even use those keys for arrows, page up/down, and home/end, opting instead to use a layer to map the vim directional hjkl to arrows etc.  Placing the keys in a cross has an implicit directional connotation, which, as you mentioned, makes sense if you're going to use them for directionally-associated functions, but it's a very counter-intuitive layout for anyone planning on using them for anything else.  With the newer configuration, the keys don't have that implicit directional connotation for anyone planning on using them for other things, but (unless I'm mistaken) will still be in a somewhat directional layout, similar to the inverted T, for anyone who does want to use them for those functions.

AcidFire, I think you're going to find this one a very divided issue with strong opinions on both sides.  I'd advise you to go with the configuration that you want, and the rest will have to wait for a swap-in board down the road.

I think you both make very good points. And there's no need to lobby for the arrows as a cluster, that'll be the first alternate cluster setup that will be released. Depending on how the campaign goes, I'm currently looking at making it the first (and very low) stretch goal since its mostly a difference in production costs.

I haven't seen any discussion of a paired numeric keypad for a while. (Really important when typing lots of numbers, and I do.) Is that going to be part of this project?

It's still in the works as an addon (and possibly stand alone(sorry inner GitS nerd couldn't help it!)) to follow the initial launch.

[CommunistWitchDr]

I think you both make very good points. And there's no need to lobby for the arrows as a cluster, that'll be the first alternate cluster setup that will be released. Depending on how the campaign goes, I'm currently looking at making it the first (and very low) stretch goal since its mostly a difference in production costs.

Really, I don't even think you would need an alternate cluster. Just some extra holes and traces in the pcb to allow for either setup. I've seen a lot of keyboards have somewhat overlapping switch areas so you can have different spacing on the bottom row, different kinds of enter key, and so on all on the same pcb.

[smferris]

Historically I've always liked having dedicated arrow keys.  I used to really like the 8-key navigation cluster on the TypeMatrix 2030, which was roughly:

PageUp	Home	Up	End
PageDown	Left	Down	Right

I'm not sure where I'd want to put that on an Axios.  It works better in a matrix array than it will on the Axios due to the finger staggering.  The new center 3x3 grids AcidFire posted about recently aren't staggered, so they'd be an option, but probably not a particularly good one since my goal would be to have three fingers hovering over the arrows and that's awkward when they're centrally located.  I may end up trying to convert myself to the layered home row approach.  The TextBlade I just ordered does that so perhaps I'll get used to it while waiting for the Axios to go into production.

[AcidFire]

Historically I've always liked having dedicated arrow keys.  I used to really like the 8-key navigation cluster on the TypeMatrix 2030, which was roughly:

PageUp	Home	Up	End
PageDown	Left	Down	Right

I'm not sure where I'd want to put that on an Axios.  It works better in a matrix array than it will on the Axios due to the finger staggering.  The new center 3x3 grids AcidFire posted about recently aren't staggered, so they'd be an option, but probably not a particularly good one since my goal would be to have three fingers hovering over the arrows and that's awkward when they're centrally located.  I may end up trying to convert myself to the layered home row approach.  The TextBlade I just ordered does that so perhaps I'll get used to it while waiting for the Axios to go into production.

Alternatively, the revamped casing for the thumb cluster (which I haven't posted yet) would allow it to sit externally so that might be an option as well (thinking about setting one up like this for myself).

Really, I don't even think you would need an alternate cluster. Just some extra holes and traces in the pcb to allow for either setup. I've seen a lot of keyboards have somewhat overlapping switch areas so you can have different spacing on the bottom row, different kinds of enter key, and so on all on the same pcb.

Have you been snooping through my PCBs? Yes, the footprints for the extra switches exist on the current boards, though I'm still struggling a bit to make them work with the RGB LEDs, I'm sure I can work out some sort of solution for production time.

Speaking of RGB, for those of you who would prefer Cherry MX stems, would you be open to the option of Gateron switches as a compromise? With a semi clear housing and from what I've seen a much more reasonable switch price, it's looking very doable to offer them along side the Matias. And from what I've read, those of you who have been wanting buckling spring may be interested in trying out the Gateron black equiv, they seem to feel quite similar (from what I've read anyways).

[Faokryn]

Alternatively, the revamped casing for the thumb cluster (which I haven't posted yet) would allow it to sit externally so that might be an option as well (thinking about setting one up like this for myself).

I have to admit I've lost track of the current layout a little bit.  This is what I'm currently working with:



That's 60 x 1u caps and 22 x 1.5u caps.  Is that correct?  If not, would you mind uploading current layout information when you have a chance?

Speaking of RGB, for those of you who would prefer Cherry MX stems, would you be open to the option of Gateron switches as a compromise? With a semi clear housing and from what I've seen a much more reasonable switch price, it's looking very doable to offer them along side the Matias. And from what I've read, those of you who have been wanting buckling spring may be interested in trying out the Gateron black equiv, they seem to feel quite similar (from what I've read anyways).

Hmm.  Now you have me wondering if I should jump on the current Massdrop for the Gateron switches    Having them ahead of time would give me time to try them out to see what I like...

[berserkfan]

Speaking of RGB, for those of you who would prefer Cherry MX stems, would you be open to the option of Gateron switches as a compromise? With a semi clear housing and from what I've seen a much more reasonable switch price, it's looking very doable to offer them along side the Matias. And from what I've read, those of you who have been wanting buckling spring may be interested in trying out the Gateron black equiv, they seem to feel quite similar (from what I've read anyways).

I strongly urge a switch moddable plate. That way we have flexibility in putting whatever stems we like inside. Not surprisingly, I'm going for the same jailhouse greens that I've been talking about perpetually.

[AcidFire]

I have to admit I've lost track of the current layout a little bit.  This is what I'm currently working with:

Show Image

That's 60 x 1u caps and 22 x 1.5u caps.  Is that correct?  If not, would you mind uploading current layout information when you have a chance?

You've got it right. For reference the current plate looks like this:

Hmm.  Now you have me wondering if I should jump on the current Massdrop for the Gateron switches  ;)  Having them ahead of time would give me time to try them out to see what I like...

You can if you like, but I believe we'll be making some switch testers available before the campaign that would include the 6 common Cherry, 5 Gateron, and 3 Matias currently available. They may also be a campaign option that would include a voucher for the CrowdSupply price (lower than the final retail) minus the cost of the switch tester. We're currently looking into it.

I strongly urge a switch moddable plate. That way we have flexibility in putting whatever stems we like inside. Not surprisingly, I'm going for the same jailhouse greens that I've been talking about perpetually.

With Cherry that definitely goes without question. If you do see me post a Cherry plate without the notches, rest assured that the production models will have them. It just happens to be easier to switch between Matias & Cherry when building layouts without the notches ;)

[DSlayerZX]

Acid, just curious, do you have an estimated date that the keyboard will be available to be crowd funded?


I am really interested in this keyboard... but my wrist has been killing me lately... I might have to get something first if it's going to take a while.  >.<

[jacobolus]

I am really interested in this keyboard... but my wrist has been killing me lately...

You should sort your wrist out ASAP, don’t wait: by all means get an Axios keyboard, but don’t just work through your pain in the mean time. RSI is no joke, and continuing to repeatedly do something that is causing you pain is a good way to give yourself a serious, possibly irreparable, injury.

You might want to read some of the threads in the ergonomics subforum here at geekhack, or start a new thread showing a picture or video of yourself typing (showing your hands and at least most of your forearm). There are many possible steps to take to improve hand comfort, including improving diet/exercise/sleep, reducing time spent on hobbies that stress your hands, working less on the computer or less overall, taking frequent breaks, changing your desk/chair setup or sitting style, repositioning or reorienting your keyboard, changing your typing style/technique, changing the logical keyboard layout (away from QWERTY) or finally of course swapping out the keyboard for something with different physical shape.

Don’t stop trying to tweak your setup and habits until the pain goes away. Pain is your body’s way of telling you something is wrong.

(Disclaimer: I’m not a doctor or physical therapist, etc. Just some guy on the internet with several friends who tried to work through pain, screwed their bodies up, and needed to take long breaks from their careers to recover or have surgery.)

[Koren]

You can if you like, but I believe we'll be making some switch testers available before the campaign that would include the 6 common Cherry, 5 Gateron, and 3 Matias currently available.

That's a really great idea. As someone that has only very limited experience with buckling springs (I'm that old T_T ) and some Cherries, I'm clueless when it comes to Matias, and never heard of Gateron...

changing the logical keyboard layout (away from QWERTY)

Plently of good advices... but I'm still not sure that Qwerty is such a pain by itself. The issue with Qwerty layout (or any 'default' layout) is probably that on typewriters, the easiest row to use is rather the upper one (shorter mechanism) than the middle one, and the layout was designed around this. Switch both upper rows and it's not a truly bad layout.

The main issue is probably the fact that the common rest position teached for such a keyboard is all fingers on the mid row. I've seen people having a resting position for the finger just one row below and have far less issues. It makes ISO layout virtually staggered, and your wrists have a far better position (angled). The only issue is that a couple of keys are a bit more farther (two "rows" up or two "rows" down, depending on the exact rest position).

[jacobolus]

Plently of good advices... but I'm still not sure that Qwerty is such a pain by itself. The issue with Qwerty layout (or any 'default' layout) is probably that on typewriters, the easiest row to use is rather the upper one (shorter mechanism) than the middle one, and the layout was designed around this. Switch both upper rows and it's not a truly bad layout. The main issue is probably the fact that the common rest position teached for such a keyboard is all fingers on the mid row. [...]

The first of these is not the issue, and the second is not the main issue; these are two problems among many.

In any case, I personally know several people, and have heard from a pile more on geekhack, who switched to Dvorak or some other logical layout and found it dramatically improved their hand comfort, despite typing each physical key with a similar style to the one they used with QWERTY. Thus, switching layouts might be one thing for someone in pain to try. (Though it’s probably the most extreme type of change to adapt to, even more effort than learning to type on a changed physical layout.)

[DSlayerZX]

Thank Jacob, for your passionate reply.

My wrist problem is mostly stem from over work than anything. (been pulling off 50-60 hours a week in office)

I have readjusted the working position already and hopefully it will help me out to reduce my wrist pain.

The office work load should die down in a week or two since I have been pushing and rejecting projects once my wrist has started to act up.... So hopefully... it should be better soon.

Though... the fact that it was acting up really raised an alarm for me and push me into actively looking for more ergonomic keyboards... hence the last comment...

[Koren]

In any case, I personally know several people, and have heard from a pile more on geekhack, who switched to Dvorak or some other logical layout and found it dramatically improved their hand comfort, despite typing each physical key with a similar style to the one they used with QWERTY.

I don't dispute the fact that the comfort will be higher if you reduce fingers motion. I'm just saying that the direct effects of a layout change on wrist pain may not be so obvious. Paying attention to wrist position even without changing the layout (and NOT using usual advices for typing, such as the one below, unless your waist is 10cm broad) is probably the most important thing to do.

More

I wouldn't be surprised that a layout with less vertical fingers motion help people maintaining a better wrist angle, explaining some seen benefits from layout. I wouldn't be surprised there's also a placebo effect in play. I've even seen people arguing Dvorak induce more strain for them than Qwerty.

I'd stay the worst thing with qwerty is not the layout, it's the way you learn to touchtype on it...

Beside that, I'm obviously in favor of better layouts than Qwerty (although I'm definitively not fan of Dvorak or Bépo, especially the alternating part)


Software that program micro-breaks seems to be useful, too...

[jacobolus]

Paying attention to wrist position even without changing the layout (and NOT using usual advices for typing, such as the one below, unless your waist is 10cm broad) is probably the most important thing to do.

I feel like we’re arguing just for the sake of it, not really disagreeing. You might want to follow the ergonomics subforum (here are some posts of mine about precisely this: 1 2 3 4 5), instead of talking about this here where we’re getting somewhat off topic.

Inre your linked image:

More

This is somewhat helpful, in particular how it stresses to avoid wrist flexion or extension, but the way the “right” image shows the forearms coming to the keyboard – parallel to each other ~8 inches apart, aligned with the keyboard rectangle – is physically impossible, and the negative slope is only relevant when the surface the keyboard is resting on is very low, not helpful in contexts where the typist doesn’t have control over the furniture. The pictures need at least a few paragraphs of explanation to give helpful advice IMO, or at least several more diagrams. The ideal is probably an explanatory video.

I'd stay the worst thing with qwerty is not the layout, it's the way you learn to touchtype on it...

The way touch typing is formally taught is mostly okay. For instance:

More

http://www.youtube.com/watch?v=fvUxizk1x3c

http://www.youtube.com/watch?v=iVBoBJ-nNAo

The way people type in practice is often awful. Other logical keyboard layouts (not to mention other physical layouts) encourage much more comfortable typing technique.

[Koren]

I feel like we’re arguing just for the sake of it, not really disagreeing.

I agree, we mostly share the same views and arguing over details, and it's indeed not the place for this kind of discussion (apologies to AcidFire and others). I'd be pleased to discuss this elsewhere with you later, but I'll stop derailing the thread further.

But just since you posted some medias, mostly the video (I love those videos from the mid XXth century, especially the science ones), I'll hide a comment below explaining the problem I have with those teachings...

More

I disagree with the most common (although not universal) finger placement/usage, such as depicted there:

I can't see how such a non-symmetric layout, with both hands going leftwards instead of inwards, can make sense...

This leads to many akward situations, such as this one:

For me, such an angle for the wrist can't be any good when you care about wrist pain.

Overall, I don't think that placing the hands perpendicular to the keyboard can be a good idea for RSI and thus, I'm not sure that "QSDF" resting position for the fingers is the best idea. It was the only possible one for old typewriters because the rows were layered, but with flat keyboards, I think it's not the best idea. Granted, I try to make an ergonomic keyboard from an ISO one, which is a strange idea, but still...

That being said, I'm out, but thanks for the discussion...

[jacobolus]

I'll hide a comment below explaining the problem I have with those teachings...

You should either make a new thread about this in the ergonomics subforum, or else do a search: there are several existing threads about this topic.

More

I personally use a fingering something along the lines of:

[smferris]

Historically I've always liked having dedicated arrow keys.  I used to really like the 8-key navigation cluster on the TypeMatrix 2030, which was roughly:

PageUp	Home	Up	End
PageDown	Left	Down	Right

I'm not sure where I'd want to put that on an Axios.  It works better in a matrix array than it will on the Axios due to the finger staggering.  The new center 3x3 grids AcidFire posted about recently aren't staggered, so they'd be an option, but probably not a particularly good one since my goal would be to have three fingers hovering over the arrows and that's awkward when they're centrally located.  I may end up trying to convert myself to the layered home row approach.  The TextBlade I just ordered does that so perhaps I'll get used to it while waiting for the Axios to go into production.

Alternatively, the revamped casing for the thumb cluster (which I haven't posted yet) would allow it to sit externally so that might be an option as well (thinking about setting one up like this for myself).

Could a thumb cluster have a 4x2 or 4x3 grid?  I was assuming it'd max out at 3x3, which would make the layout much more awkward.  I found it really useful to have PageUp and PageDown stacked on top of each other just to the left of the left arrow, so that I could easily reach them with my index finger while keeping the rest of my hand at a home position over the arrows.

[AcidFire]

I will agree 100%, do whatever you need to do to deal with your wrist pain, your health absolutely comes first.

As for a 4x2 or 4x3 cluster, 4x2 with 1u keys *might* be possible, but a 4x3 would require a larger housing. Originally there was space for this, however in order to keep the distance between the keys of the main cluster and the thumb to an absolute minimum, we went with a smaller housing to match the area available. This was part asthetics but mostly due to manfacturing & inventory management, a larger thumb cluster means there is a higher molding cost (for two unique cavities) since each side now has it's own MOQ. Plus, inventory management/space to keep track of the additional part since now they need to be tracked and verified seperately when building kits.. It may not seem like much, but with everything else that goes into this it adds up very quickly.

[smferris]

As for a 4x2 or 4x3 cluster, 4x2 with 1u keys *might* be possible, but a 4x3 would require a larger housing. Originally there was space for this, however in order to keep the distance between the keys of the main cluster and the thumb to an absolute minimum, we went with a smaller housing to match the area available. This was part asthetics but mostly due to manfacturing & inventory management, a larger thumb cluster means there is a higher molding cost (for two unique cavities) since each side now has it's own MOQ. Plus, inventory management/space to keep track of the additional part since now they need to be tracked and verified seperately when building kits.. It may not seem like much, but with everything else that goes into this it adds up very quickly.

4x2 1u is really what I'd be looking for in a navigation cluster.  I don't know what I'd do with the extra keys if it was 4x3.  I wasn't expecting it to even be able to handle 4x2 keys, so it's good to know it might be possible.  When do you think we'd know if 4x2 was workable or not?

Overall I think a smaller thumb cluster makes sense, at least for it's intended purpose.  Tacking a navigation cluster on elsewhere would be nice if possible, but at that point I'd be trying to reuse a part for something it wasn't designed for.

[AcidFire]

I suppose I should have been clear/clarified, if you're looking for two rows of four, that won't fit without a larger thumb cluster. 4 rows x 2 columns, however is definitely doable.

[jeep]

Any update on how long you think this will take? Also, has the multiplxd forum been pretty much abandoned for now?

[richardkemp]

I hate to be that guy, but I'd also love an eta. I need a second ergo board since I currently commute with my ergodox, and I'd much rather wait a little while for an axios than give in and get the ergodox update currently going on massdrop. That said, if eta is more like a year than a few months... compromise time. Can you give us an estimate acidfire? 

[rsadek]

I hate to be that guy, but I'd also love an eta. I need a second ergo board since I currently commute with my ergodox, and I'd much rather wait a little while for an axios than give in and get the ergodox update currently going on massdrop. That said, if eta is more like a year than a few months... compromise time. Can you give us an estimate acidfire? 

I'm in a similar boat. My plan is to hedge: add 2nd Ergodox, then sell one once the axios comes. For a small cost this will avoid frustrations such as timeline slips,  damaging or losing Ergodox during commute, etc.
-R

[jacobolus]

That said, if eta is more like a year than a few months... compromise time.

As a general rule, if you need something in the near future, just buy something that already exists, instead of waiting on an unfinished version 1 hardware project (same story whether it’s a 3d printer, a blender, a chair, a shoe, whatever...). You can always resell whatever you get, when the new hardware finally arrives: the ergodox or kinesis advantage or whatever existing product is going to hold most of its value for at least a year or two.

It’s wonderful to support (via crowd funding or pre orders or design contributions ...) new hardware projects, but go into them with an expectation that the final ship date will always be at least 6 months later than you expect [or judging from the Kickstarters I’ve been involved in, often at least a year behind the intended schedule]. Then if a project meets its schedule goals you can be excited and happy, but you won’t feel as bad about delays. If you need and expect to get something in your hands by a particular date and you pre-order a product still in development, you’ll be really disappointed by every tiny schedule slip, and end up grumpy instead of excited about the new thing.

[clickclack123]

As a general rule, if you need something in the near future, just buy something that already exists, instead of waiting on an unfinished version 1 hardware project (same story whether it’s a 3d printer, a blender, a chair, a shoe, whatever...). You can always resell whatever you get, when the new hardware finally arrives: the ergodox or kinesis advantage or whatever existing product is going to hold most of its value for at least a year or two.

I think you are definitely correct about ths, but just a comment, the Infinity Ergodox sounds like it's a bit of a work in progress as well. They're still investigating whether stabilizers will be included in the drop, there are no photos of the lcd working, also it doesn't sound to me like the backlighting will have fw support in the first instance.

I do think that the basics will work when it ships though... I should think that though, I bought one!!

They did say in the description that NKRO and the lcd would work though.

[jacobolus]

I think you are definitely correct about ths, but just a comment, the Infinity Ergodox sounds like it's a bit of a work in progress as well.

Yep. Ergodoxes in general take at least a few months between order and ship date. If you need something within the next month or two, buy a retail product, or build it yourself.

[dsmitify]

I did the same as clickclack123. Ordered an Infinity Ergodox while waiting for the Axios