Google’s project Ara, aimed at creating a smart phone consisting of easily interchanged modules, is officially shelved.
Ara should have been a research project, and not a productization attempt. As a product, it was always questionable: it’s hard to figure out what could make customers wait in line for these phones (“I can change a camera, uhuh, OK, and?”), and it’s not clear what would make manufacturers want to produce them (“Let me see if I get this straight, I now get to sell components with lower profit margins, rather than integrated devices with higher ones — I should do this why, exactly?”).
For a research project, seamless modularization would have been an intriguing and potentially transformative goal. Modularity is cool. Done right, modularization of a core service makes it easily accessible — you can then unleash worlds’ engineers on building whatever-their-hearts-desire on top of the core functionality. We’ve seen this with the Internet: define interchangeable layers, and let the world loose on it — the results exceed anyone’s wildest expectations. Facebook and Youtube wouldn’t have been possible without the humble OSI stack. But this is a vision for a longer term than what is possible if you want a near-term product out of it. You have to aim to kick-start a different era of technology, rather than to sell a gadget this Christmas season.
Ara being a corporate productization effort, none of learnings from its failures will be ever shared — and that’s a real loss for the technology community. Ara failed in part because the performance story for a modular phone just does not add up. The current phones are co-designed through and through; it takes person-years of engineering effort to fine-tune this co-design to get a phone to its current level of performance. Modularization incurs a performance penalty always; currently this penalty, for smart phones, is prohibitive. Developing a fundamental understanding of the nature of this penalty, fully understanding the performance bottlenecks, establishing a specific direction for which technology breakthroughs are needed to make modularization a possibility and then working on those breakthroughs — that would have had a real impact on the world, even if it failed to produce the full desired outcome. Ara should have been a research project that the community could learn from: then even if the project ultimately failed, we would have learned the limits, and would have developed new ways of looking at related problems.