ARKit Fail? Nah, the Technology Is Great, but It Might Not Be Ready for Engineers Yet

ARKit already has its useful points, at least from an entertainment perspective. Using Holo, I’m able to bring a trained lion into my kitchen, and snap photographs of one of my housemates interacting with her. That’s definitely a win, as opposed to ARKit fail.

There are also some very cool games that bring the action into your space. ARise, for example, puts a three-dimensional magical puzzle on your floor or table, so you can move around it and try to solve the quests. Others, like Warhammer 40,000: Freeblade, allow you to bring your mechanical war machines to life in your room or yard. Our own Trevor Sheridan has checked many such games out, and provided you with a list of some of the top ARKit-enabled games.

As great as AR might be for entertainment, I’m not yet convinced it’s ready for prime time outside of that niche. We aren’t talking complete ARKit fail, but it has some limitations. Based on the numerous apps I’ve experimented with over the past few days, ARKit is just not quite ready for prime time. I’ve tested apps that allow you to visualize furniture in your room, make use of an AR tape measure, calculate distances, and determine angles.

Some functions work really well, if you know how the technology works. ARKit needs to glom onto a flat surface, for example, and make calculations of the physical size of that surface. Linear measurements, like with an AR tape measure, are fairly easy to accomplish. I can take measurements, verify them with a physical tape measure, and they’ll come out fairly close, if not spot on.

Beyond that, things start to fall apart. For example, an otherwise excellent app that offers a feature to measure the distance from one point to another tends to fail miserably if the perspective or lighting is off. I watched a 20-foot measurement get reported as six feet, changing to the correct distance when I adjusted the light in the hallway. That’s a problem that many users might not figure out.

Another problem lies in ARKit accurately determining where the floor is. Depending on your carpet and lighting, The IKEA Place app, for example, might have you viewing furniture floating a few inches off the floor instead of resting on solid ground.

Now, try measuring angles. This turns into an absolute nightmare, and almost had me convinced of ARKit fail. Then I started to really think about what needed to go into this function working.

First, you need to have extremely steady hands (or use a tripod). Next, you have to keep your perspective just so.

Skew your camera few by just a few degrees along the Z axis, and the ARKit loses its ability to offer you an accurate angle. What should read as a perfect 90-degree angle (I checked with a T-square) comes back as 120 degrees in one measurement, 110 degrees in another. That’s a big problem for anybody who might use ARKit-fueled apps for construction or engineering purposes.

There’s also the problem of how the apps are developed and tested. An app that allows you to measure the distance from one point to another allows the text of the measurement to follow along with the pointer, keeping in perfect proportion to the ending spot. That’s all well and good until you’re measuring a distance more than 10 or 15 feet away, and then the text becomes too small to read. In fairness, I’ve spoken with that app’s developer, and expect that particular issue to be resolved in the next update.

Before you get too worked up and start ringing death knells for Apple in general or ARKit in particular, remember this. The technology is still new and evolving. I fully expect that a few months of testing and tweaking will result in much better apps. For now, though, ARKit is great for entertainment and casual use, but it isn’t very useful for more serious purposes. To answer my earlier question, ARKit isn’t a failure. It’s just in need of a bit more incubation and maturation.