Matias Duarte, Head of Design at Android, has recently held an interview on software design during Accel Design Conference 2014 underlining the need for a shift in software design approach from separate apps made for different devices to one app for multiple screens.
Addressing software designers, product managers and developers, Duarte made a pretty bold statement during this interview: “Mobile as a concept is dead, mobile is over.” He certainly did not mean that mobile devices are dead, but designing apps for mobile devices as separate applications by themselves is the wrong path to follow. He suggests instead thinking in terms of “Screens”.
Paraphrasing Duarte, mobile was seen as a different category when it was constrained, when the bandwidth and processing capabilities were limited, but now, for most people, any experience one can imagine can be meaningfully expressed on any screen size. This means that designers now need to stop thinking about mobile as a distinct category, a separate group with a separate set of metrics, etc.
Duarte advises to invest into the screens that people are interested in, but not entirely in the 5 to 7-inch ones. Also, one should not create different products for different screen form factors, but rather develop one product with a unified design approach that spans multiples screens: the desktop one, the smartphone, the in-car solution, the smartwatch. People should be able to continue to use a certain program while switching between all these different screens throughout the day. This does not pertain only to the visual appearance of a product, but also to its feature set. If one makes an app for taking a taxi ride, it is not ok if the app works on a 7-inch screen but it does not work on a 2-inch one.
While Duarte advises product managers and designers to think in terms of screens, he admits that the technology is currently young, and there are some issues, such as different technology stacks on certain screens compared to others, as it is the case with Chrome OS on laptop and Android on smartphone, but the users should not suffer because of that.
As a side note, without being explicit, Duarte nodded that at some point Chrome OS and Android will have to converge. Actually, all Google platforms, including TV and in-car will have to converge, without providing any details on how this will be achieved.
Duarte’s basic idea is having one product across multiple screens rather than different apps designed for different devices. “Design for users not for devices.”
When it came to the native vs. web debate, Duarte noticed the following:
- This preference for native instead of web is fueled by technological “fanaticism”, by the preference for what people know, what they can understand.
- There are a lot of attributes about the web, about the way services and experiences can be delivered, that are amazing, and they are not present in native apps.
- Developers should go past the present concept of apps, because users don’t want to install an app, they just want content.
- Currently, on the web one can get hot linking, indexability, efemerability – getting the content one is looking for instead of having to deal with an app, having to install and manage it, but on the native side one gets better performance, access to sensors, etc. The future needs to include the benefits of both. It is the job of OS and software makers to work on eliminating the current confusion introduced by having different platforms for different devices and native vs. web while heading towards a future where the technology beneath a product is not important but rather the content provided.
- “Web and [native] apps are both wrong, they both have to die.”
Regarding the design philosophy behind Android Wear, Duarte underlined the importance of significantly shrinking the time needed to interact with a smartwatch rather than having to stare to such a device, so the user can be more available for real-world interactions.
As a biographical note, Duarte was also deeply in the design of T-Mobile Sidekick, the Helio Ocean, and Palm's webOS.