I've said many times that I - and many other Treg testers - have found through their experience and discussions with Mark (and each other) that the long delay in shipping is understandable. That the TextBlade is far more complicated than people (including the Treg testers) realized before getting one.
Soooo, being bored, I decided to do something I've done a few times before. You can blame this on dpom who reminded me about one of these projects from well over a year ago. Anyway, I'm wading through the reports of Treg testers here, picking out specific comments on the issue of shipping delays and whether it is worth it.
I don't claim to catch all the comments - most of these posts they made are long so I was skimming them. And I focused mostly on first posts in threads, though I usually scanned the rest.
The following will list the names of each tester I've checked so far. So have no comments on this matter that I saw. Sometimes this may be because they made such comments, but in posts earlier than I've gotten to so far. The order isn't consistent with time of posting, but generally works from most recent to older. At least one goes back to the end of June so I have a lot more to go through.
VERY long post coming up:
I've got a strong new understanding of why WayTools doesn't "just ship." Why "good enough" really is not as good as what I'll call "a vanishingly fluid disappearing act!"
I also gained an understanding and appreciation of the enormity of the task of creating this bitty thing. It takes a lot of machine discrimination to make the proper difference between
- a ideal keystroke and the more typical keystrokes that a gazillion different humans will make. They maybe are
- struck off center
- struck with fingernail instead of the pad of the finger
- struck from an angle
- bridge two different keys with the desired key getting the stronger numeric capacitance score
- bridge two keys with both keys getting very similar intensity of capacitance readings.
This latter condition is one that has required developing a process for machine learning/AI discrimination to develop the rule that says, "This particular 'human' really wanted an i in this location instead of the k that he also pressed."
And I am convinced that this level of problem is something that is reasonable to have been discovered only after the first full production run.
There may be a tendency for some people to trivialize the accomplishments that the TB represents due to either its small size (how complicated can a little thing like that be?) or to comparing it to other keyboards. Those of you with hands-on experience know better. This device is a HUGE breakthrough in keyboard design and manufacture.
WayTools delay of the general release has been due both to the need to overcome real technological hurdles and to the extremely challenging nature of those hurdles.
My sense about the delay is also that circumstances have conspired to deal WayTools some cards from the bottom of the deck.
The technology is extraordinarily complex. I have no trouble believing this is the most challenging keyboard ever built, by an order of magnitude. As a developer who’s worked on software and hardware products for my entire career, I have some standing to tell you that.
• Multi-touch capacitance with significant key travel in a low-power device creates an unusual difficulty with signal-to-noise-ratio. The device does not use high voltages like a touch screen to boost the signal. Instead it shields its electronics so it can use low voltages, and applies advanced signal processing and machine intelligence to detect the movement of your finger and to infer your intent (which key you want to strike) even when your finger is actually closer to another key.
I'm also very confident that it was not possible to deliver this earlier than now and that Waytools have not dragged their heals at all on this. All gut feel, interpretation, reading between the lines, experience, and reading the forum for the past 2 years. I can't wait for this keyboard to hit general release, but I also appreciate and respect their decisions to not release in the past. I for one am glad that they held off.
I've been working with WayTools to try and hunt down these problems, but the fact is that a lot of these would be crippling to general release right now. And sure, I'm a weird edge case in how I can't get along with iOS and all the custom tweaking I'm trying to do, but it's still enough of an edge case that I can understand WayTools wanting to make sure that these problems don't cut off general release at the knees
The delays are unfortunate, but obviously necessary.
When WayTools started selling the TB they must have felt the hard work had been done and what was left was fairly simple: turning the working prototype into a reliable and scalable manufacturing process. Not only did that turn out to be harder than they thought, when they started testing the end-result in real-life situations their passion for excellence began to work against them. Because the TB was conceived as something much more than just a keyboard they couldn't allow it to go to end-users functioning as just that. It had to be immaculate on many more levels, to meet their own excitement and espectations. There were also no pre-existing standards to measure success by; nothing comparable exists on the market today. So it took much longer than they would have expected at the start, because while they were ironing out kinks they also had to create the very standards to measure progress by and the equipment and processes to test against those standards.
I know it's been mentioned, but this thing is crazy high tech. I honestly wish they'd post some of the stuff he talked about here somewhere. Talking about how sensitive the capacitive sensors needed to be, the work that went into the angle of the blades, the thought that went into where things were placed, the software behind it all is just nuts.
This keyboard is worth the wait even if you had to wait an additional year.
My discussion with Mark truly showed his passion for what they are debpveloping and the delays are understanble given the complex nature of the tech behind the TB. There are so many hew functionalities this device has, most of which were actually unknown to me.
There are a lot of little things about the TextBlade that have been carefully thought through. I wasn't able to appreciate most of these things in the first week; however, I believe that this no-compromises approach* to the product will serve WayTools well in the long-term. In short, given what I've seen as a member of TREG, I'm more willing to trust WayTools' decision about when the product is ready for general release.