Note what is happening in the car industry where an increasing number of people are thinking of the personal smartphone as the CPU with car peripherals such as sensors, displays all connected to it. Why do we replicate these on the phone? The core phone should have the processing power and basic peripherals, anything else should be a pluggable accessory.
I say that's what people used to say in the early days of computers, now we can interchange displays, keyboards, mouses, HDDs, USB peripherals etc. on our PCs without a second thought... the same will eventually happen with phones....
Given the cellular service carrier is pushing devices to consumer and giving out discount every 2 years, I really doubtful consumers will be interested in upgrade modules instead of upgrading the hardware. A modular design is a great concept and will ultimately achieve a great product. However, we have to remember extra packaging will be needed to achieve the goal. It means the "wanted" device is going to be bulky compared to a one for all devices.
Having said all these, I don't discount the value of the modular device. It open a new door of business - module development. It provides a way to test the market to confirm whether a particular module "feature" is getting enough attention from the market. Later, mobile manufacturer may consider putting in as a standard feature. There are definitely different niches to be addressed. For example, devices can be put together in different way to monitor different patiences and to address the need of them.
Question is what is the different between a docking device to your iPhone than a modular phone.
It is surprising that the 10pin I/Fs are located on the top/bottom surfaces of the frame rather than at the vertical ridges separating each location. These connectors also seem to be awfully (and relatively) large for the minimal number of lines. I would have expected a paradigm shift via mostly optical interconnectivity but we may not be there yet!
The plug-in modules keep reminding me of pieces of gum and the phone itself is the equivalent of a package of Chiclets!
I was skeptical when I first heard of Ara, but it appears that the team has made huge strides. Whether this will truly have mass market appeal or just fill a niche depends on so many factors, it's hard to say at this point. Certainly module support from third parties will be essential, and the overall price with modules that provide similar functionality to a high end smartphone needs to be comparable. Those who want to add more features & functions, or higher performance functions -- better cameras, more cameras, etc. -- will of course be willing to pay more.
I see this as more useful for fast prototyping than for building an actual phone. Let's say you have a new gyro and maybe a heartrate sensor you want to put in your next gen phone. You could build up your phone out of the various pieces, then give it to software to start writing code and getting working, while you design and build the real, non-modular phone. The software will be ready before the phone is even built.
Likewise, a company could develop a new sensor or graphics chip, and develop a module for the Ara, so companies could immediately start prototyping with their sensor.
I love the concept behind this project, but I don't see these showing up as mass market consumer items on cellular networks. What I see is tremendous potential as a prototyping platform and low-volume manufacturing technology for the next generation of IoT devices. It starts with a low-power but powerful processing core, adds an arbitrary and flexible set of sensors, and tops it off with a selection of connectivity options. It can be a phone if it needs to be, but if it doesn't then just eject that baggage. The M2M guys have been working with cellular modules to get similar capabilities, but this wraps it up in what could be a very usable package. Am I off-base here?