So yeah, there are definitely a few major problems in the data glove space right now.

First, human fingers turn out to be pretty strong in comparison to the types of force-generating things we could fit in a human-hand-sized box. We looked into a few different ideas, but there is a big trade-off between actuator size, amount of force the actuator can create, and the speed at which it can actuate. Unfortunately anything that is small enough and strong enough is probably too slow.

Second, you just can’t get a glove. I’ve been following the space for a while, and the article you linked even mentioned a few options I hadn’t even heard of. There are a lot of glove projects out there, and you can’t really buy any of them. I think the situation in the market right now is kind of ridiculous. Building gloves is not really that hard. I built my first prototype two years ago, in a weekend. No, a prototype isn’t a consumer grade product, but just how long does it take to go to market, really?

For example, the Dextra Dexmo is probably the only data glove project right now that has anything realistically close to a good trade-off on the size/strength/speed spectrum. But you also can’t get it. I’m also a little concerned that the back of the hand is not as free clearance as we might tend to think, but I’m willing to be proven wrong there, if I could actually get my hands on one.

There was another glove project a few years ago that used air bladders for force feedback. Air bladders are certainly small and you’d be able to get a fair amount of force out of them… but they are incredibly slow. There’s probably a reason nobody has heard from that project in a while.

Most of the design features for this glove were dictated by the short amount of time we had to put into it, coupled with the desire to make something that actually worked and wasn’t a disappointment to anyone who used it for more than a minute. I keep going back to thinking about the Leap Motion and how amazed I was at it for, like, a day, and then got super depressed at how impossible it was to develop anything good with it.

I knew we needed Lighthouse to have rock-solid gross positional tracking. I knew we could make flex sensors for the fingers work for the context. And I knew the only sort of feedback we’d be able to build that would work reliably would be vibrating motors. We didn’t have a lot of time to build it. I started the Bluetooth code basically the end of November, during Thanksgiving break while taking Valve’s HDK training course. Erik started playing with physical construction ideas end of December during Christmas break. We had almost no work on it during January, trying to pay bills around the office. We basically put this all together in about a month’s total worth of work.

Ideally, I’d like to have Vive pucks instead of the Vive controller, but Valve doesn’t seem to know anything about us, I don’t know why. We’ve taken their HDK class, we’ve done lots of conferences, we’ve applied to everything they’ve put out, and they routinely give the first dibs of hardware to people more interested in playing games than building things. If we get the Vive pucks anytime soon (one is supposedly on its way), we’ll be redoing this glove to have the main circuitry either in the back of the hand or under the wrist, with the puck being like a large watch face on the back of the wrist.

So yeah, data gloves have problems. They are mostly vaporware. Which is ridiculous, they’re actually pretty easy to build. Our design isn’t ideal, but it’s a start and I think any data glove is better than no data glove. It primarily gives us something to start designing software around.