If this ever takes off, the real upside may not be the modularity itself, as that'll have such a negative impact on efficiency and form factor and the user interface. But assuming they get it into a somewhat tolerable range, the really interesting part will be analyzing the which modules get developed, which modules get used the most, and which ones don't really get used all that much.
From that, the ultimate efficienty and usable smartphone can be developed based on what users really do or choose to have and use on their device. From features to form factor. This analysis really fits with Google's 'analyze this' approach.
I really can't see this going anywhere. Google is great about trying out interesting concepts and seeing where they go. I wouldn't put any weight in their investment (remember google wave? Feedburner?).
It is a cool idea and a fun concept, but the "cons" column is so much longer than the "pros" column that it is hard to even fathom this concept going into production.
Agreed, this may be something down the line, but for now its more of a curiosity, more of an interesting little investment from Google. They are a big enough company they can spread out and try different things with minimal risk. I wouldn't place too much of an emphasis on this latest intiiative from Google.
Another question to ask is "why would a vendor put the effort into creating a module for Ara?" or rather "who will be making these modules?"
A vendor can sell a camera to many many product creators. However, if the vendor creates a camera module for Ara, it can only be used on the Ara market. Is google to going to make these extra modules? They're the only ones with incentive.
I must be the odd man out in this conversation but I like the whole concept of modularity: Heathkits were cool. Separates in the audio industry was even better. I helped rebuild a VW dune buggy and learned immensely from the process. Building your own computer from components/hw was satisfying. Gamers created a whole cottage industry and benefits of high end computer parts are still with us but almost on its last leg. Even Arduino lovers can't be wrong. Along the same lines, instead of the current throw-away smartphones that fill entire dumps, being able to upgrade incrementally, a la Project Ara, sounds nice but something tells me it will not get to a maturation state. And too bad for that!
Microsoft and Intel set down the specification for hardware for years. Companies like Dell, Lenovo and all the other PC clone makers were building and packaging reference designs. Not to say that is not a difficult and complex buisness to manage.
This sounds very similar. Here are the starting building blocks for Android prototypes. Now that most of the R and D is done, package it up and sell it cheap. That spreads the Android app store, and keeps the customer happy. No one wants a black eye in the market because new Android software is running on an underpowered, incompatible hardware platform.
@dougwithau, that was exactly my initial thought. How would this be different from a so-called "reference design" chip companies and platform vendors like Microsoft have been offering for decades?
Well, as a starter, hot swappable "module" approach might appeal to hobbyists.
But then, it's hard to imagine that this is being pitched at hardware suppliers (ODMs) who have never made smartphones before but want to make a ton of it. For that, companies like MediaTek, Spreadtrum and others already offer a very comprehensive reference design (or what they call "turnkey systems.")
One thing is clear, though. The fact Google is experimenting with this modular approach on the Android smartphone platform goes to show how mature smartphones have become as a consumer product.
Junko, it is a bit odd that Google has focused this solely toward smartphones. When you factor in certifications and software compatibility, this just isn't a viable solution for a usable smartphone. However, I would agree that it does make an innovative development platform for IoT and embedded applications.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...