AndroidTO: The Droids I’m Looking For

Yesterday, I spent the day at AndroidTO. I attended this event last year and it seemed like a nice, little conference with some concentrated discussion about the Android Platform — I made sure to watch for the event coming around this year, too.

It’s a smaller event, but still hosted several hundred people. And this year, Intelliware was one of the sponsors and it was fun to have the booth out again.

Here are some recaps from some of the topics that I caught.

The morning keynote was by Akshay Agarwal from ARM. His talk was called Mobile in the Modern Age: Connecting Intelligence. The session was quite technical (and a lot of it might have been lost on the business and design types in attendance) and addressed several new developments in the ARM chips. One of the big topics I took away from that session was the way mobile is driving a need for higher-end graphics capabilities separate from, say, gaming.

Another point that was made was the way Android was driving convergence across a lot of different spaces — phones, cars, TVs, etc.

My first non-keynote session was part of the developer stream. It was a presentation by Ian Darwin (author of the Android Cookbook) about developing for Android using technologies other than Java.

  1. App Inventor
  2. SL4A (a general scripting tool with support for Python, Perl and others)
  3. Ruboto (a Ruby tool)
  4. HTML5 (via Cordova or Cocos2D)
  5. Other Languages like C++, Scheme and Clojure)

Certainly HTML5/JavaScript is a common programming paradigm, and C or C++ using JNI is a well-understood possibility, but many of these were quite new to me. I’m suspicious of App Inventor, I must confess. It seems like another one of these “programming by dragging and dropping” tools that I’ve never seen work for more than the most trivial applications.

SL4A sounded like the most ambitious of the tools mentioned — it’s trying to let you write code in a scripting language. Darwin responded to a performance question by suggesting that the performance implication is not zero but probably not noticeable. The real downside to my ears is that there’s an annoying script editing cycle. I could see that getting real old, real fast.

It feels like HTML5 is still the most viable alternative, but it’s interesting to see that other possibilities exist.

For my next session, I popped over the the UI/UX stream to see a talk called “How Important is Design/UX In the Mobile World?” I confess that this talk felt the least coherent to me. Adam Shutsa from 500px talked about the importance of engagement — this felt to me like a tie in to all the talk people are making about the rise of systems of engagement (in contrast to record-keeping systems that have been the lifeblood of IT for the past five decades or so). But Shutsa didn’t draw that connection.

He asserted that engaging apps needed three things: familiarity, trust, and comfort. He talked about each of these points, but I didn’t ever feel like I got the a-ha moment out of them. I responded more to points he was making about consistency, and how that manifests in two different consistency needs (which sometimes work in opposition to each other): the need for an app to be consistent with the organization’s brand (so that we’d expect the 500px app for iPhone and Android to have a certain degree of consistency) and the need for an app to be consistent with the device’s OS (so that we’d expect the Android app to behave like other Android apps).

Shutsa did acknowledge that mobile apps have very different engagement patterns, so he wouldn’t expect the mobile app to bend over backwards to be too consistent with the web app. And finally, he talked about the three laws of robotics goals of 500px’s UX: display large photos, show activity info and give people the ability to “Like” and “Favourite” photos.

I’m not a UX person, and I don’t play one on TV and I’m afraid I may have failed to appreciate a lot of the wisdom of this session. Sad.

After lunch, I caught a session about Lagoa in the Gaming stream. Lagoa is a 3D modelling tool and rendering service built for the HTML5/Software-as-a-service generation. And, in that respect, it’s a pretty impressive offering.

Although it was in the gaming stream, speaker Thiago Costa acknowledged that a lot of his existing customers are companies like shoe and car manufacturers. That those types of companies are interested in letting people have fun, game-like experiences with their products. Customers can, for example, see what their car would look like with all of their customizations and in the paint colour of their choice.

And this ties in, again, with the theme of mobile and engagement. These companies have clearly realized that giving people a reason to “play” with a product on a mobile device leads to increased engagement with the brand. It’s weird that this was consistently part of the subtext, rather than the actual text, though.

It wasn’t quite the session I was expecting, but it is one of the talks I keep thinking back upon and thinking, “Hm. That’s interesting.”

And then I went back over to the development stream to catch a session on what Android will look like in 2016. There were a lot of interesting claims being made by Shidan Gouran of Home Jinni Inc. And only time will tell if they come true.

He had a wide variety of opinions: “Bluetooth will be important.” “NFC won’t go anywhere.” “Devices will be 4x more energy efficient.” “By 2016, indoor maps (such as, for example, shopping malls) will be prominent.”

One of the most interesting ideas he talked about was that, by 2016, our mobile devices will be always on, always sensing and engaged in a “cognitive cycle” of taking in information through sensors, analyzing that information, reasoning from that analysis (so that it can provide contextual answers to user requests) and adapting that reasoning process over time. And because of improvements to natural language processing, we’ll be able to command our devices with basic speech. Of course, this all makes me think of SkyNet.

He also asserted that the camera will emerge as the single most impressive sensor on the device. All-in-all, an interesting talk; I did wonder, as I often do, what are the sources for his predictions? Gut feel? Extrapolation from industry data? Future knowledge obtained from a captured time traveller? Without any sources, who can say.

And then, finally, the closing keynote was given by Tumblr’s Chris Haseman. Haseman did the closing keynote last year, and he was excellent. My worry, of course, was that it’d just be a repeat of last year’s session. But I shouldn’t have worried. This year his agenda was:

  • Developing: Patterns for developing and releasing Android software;
  • Designing: Tricks for designing and building beautiful software; and
  • Marketing: Strategies they use to engage with users and protect the app store rating

This was, in my mind, the best session of the day. Just about everything Chris said was interesting, seemed clearly backed by lived experience, and was delivered in an entertaining and engaging way (even when he was nearing the end and racing through the last few slides).

There are way too many points for me to do justice to the overall talk, but here are some of the points that I enjoyed the most:

  • Trust is the most important currency that developers have. That means that developers and their managers need to learn helpful strategies to schedule meetings in ways that allow developers uninterrupted development time and how to estimate work that isn’t just assuming the best case. Over-padding can erode trust just the same as under-estimation.
  • Design is incredibly important — and so is pixel-perfection. Haseman said that users wont’ be able to tell you that a margin alignment is what’s bugging them, but they’ll feel that something’s off and they won’t like your app. Haseman says that he’d often overlay the designer’s image on a screenshot of the app, and fiddle with the opacity to try to spot differences between the design and the implementation.
  • It’s important to validate assumptions, and in order to do that, you need numbers, metrics, etc. You need to check to see if the change you implemented lead to positive user change. Also: “Numbers are your only weapon against bad ideas, especially your own.”
  • You need to engage your users. Respond to all feedback through the app store, even the most mean-spirited 1-point ratings.
  • You also need to ask your users to rate you. Don’t just hope they’ll rate you: ask. Don’t be obnoxious (although I liked his fake ratings-nag pop-up: “Give us 5 stars right now or Christmas is cancelled!”), but if you don’t ask, the users won’t think about it. Haseman says his rating went from 3.6 to 4.5 overnight just by implementing a plea for ratings.

A great, great final session.

A few less-than-happy-making points:

  • While the presentation and seating facilities in the theatre are great, the “vendor alley” isn’t quite as conducive to wandering around as I would have liked.
  • I know that they’re still a small, new conference, but it’s a pity that there were zero women presenters. I like you, AndroidTO: don’t become a brogrammer environment!
  • I hate hate hate the 8:30 AM start time. I hate it even more that they don’t actually start at 8:30, but are in the neighbourhood of 20 minutes late starting.
  • It’s a real pity that Google has no presence, there.
  • The wifi was not happy with that many people.

But even with those minor negatives, I look forward to AndroidTO 2014!

It's only fair to share...
Share on FacebookGoogle+Tweet about this on TwitterShare on LinkedIn