,
22 tweets,
5 min read
Read on Twitter
Next session: talk on designing AR in Maps. #io19
Jeremy Pack discusses navigation challenges in crowded places: GPS miscalculation and triangulation. Reminds me of Tango sessions on world mapping/location in the past)
Pack has a pretty critical take on GPS/compass inaccuracies in phones and their challenges. Problems in inertial odometry. AR visual mapping corrects it. Maps AR uses maps to calculate and predict against camera info 
Trees throw off recognition. Algorithms learn to recognize persistent things vs those that change over time. (Light/dark too)
(I love this conversation because it’s discussing challenges in how to have AR be persistent over time.)
(AR Maps, to me, almost feels like the guinea pig for exploring these challenges in more apps)
Looks like AR in Maps could help solve for evolving better direction suggestions (maybe also when AR isn’t used, too- could AR help build better models?)
Rachel Inman now on stage to talk about user experience (UX) design
I want to live in a multi-dimensional city #io19
Early AR maps experiment used blue line on ground- didn’t work. “People felt compelled to walk right on blue line.”
Other Map prototypes including last year’s fox. What happened to it? Fox experience was hard to get right. People expected fox to be smarter, avoid shortcuts, or lead to interesting things. Too many expectations on AR character. Still prototyping and testing fox for now.
Another test: following stream of particles that would show path. But people hated it: people described it as following “trash.”
Settled for 2D map below for easiest way to get a heads-up.
Principles per Google on AR: 1-embrace the sense of immediacy. Hard to manage user attention. Focus on one thing at a time, not flooding the zone. Near vs far, visible vs not.
Principle 2: grounding and glancability. Be clear but stand out from real world. Precise goal to NOT blend with environment, so can be spotted.
Principle 3: leverage the familiar. AR makes up “small percentage” of experience. Represented by blue signs, red dots, street signs, as normal.
Final principle per Joanna Kim, product manager: originally seen as an AR mode done totally w camera. But shifted away from that- everyone staring at screen. And people started to glue to screens. Didn’t want that.
Google explored ways to discourage staring at phone. Currently blocks AR when you start walking. That principle may change over time. It’s trying to solve a problem with walking/AR and attention. (Also saves battery, drain while in AR hasn’t been solved)
Final principle is “Keep AR moments short and assistive.” Hmm, interesting. Similar to idea of short interactions on watches. #io19
Didn’t want any headline that said “Google Maps launches useless AR feature,” Kim says. Wanted to slow deploy and test with local guides, first. “We clearly have a lot of work to do.”
What’s next: working on assisting better at night, expand street view coverage in more places around the world, localize users faster and more reliably. Pixel launch another step, but still exploring how it works. #io19
That’s it. Sorry for tweet explosion. I find this stuff fascinating.
