Android OS CarPC?

- - posted in Cars, Linux | Comments

Generally speaking, I don’t actively participate in the weekly #carchat on Twitter, but I do occasionally glean a few points as I watch it stream across my Twitter timeline.

Take for instance this week’s #carchat where I saw @banovsky say the following. [twitter_status_by_id id=“13882824556”]

Now, I’ve built a CarPC and have some familiarity with the topic, and on the surface this seemed like a REALLY good idea, so I excitedly replied. [twitter_status_by_id id=“13883363838”]

And apparently @bprosperi agreed; [twitter_status_by_id id=“13883473947”]

Admittedly however, my response was not based on any particular facts or prior research, it just sounded like a novel idea. So when I was challenged, I honestly didn’t have anything particularly intelligent to say. [twitter_status_by_id id=“13883392253”]

[twitter_status_by_id id=“13883614504”]

That was pretty much it for that conversation, I was running out the door to head home from work, and there didn’t seem to be a lot of activity on that conversation topic. But the brief exchange did get me thinking. I’m going to be tackling the challenge of updating my CarPC and installing it in the Caddy at some point in the future, and it might make sense to consider my options in the Mobile OS arena. With that, I decided to refine my understanding a bit.

What I found, was actually quite disappointing! It seems that a lot of folks in the car computing community have given this topic a lot of consideration, starting all the way back when Android was first announced. The first several links of the search I did were all links back to the forums, and I observed a few common threads.


Android is hyper focused on small mobile devices (read teeny screens), so while it’s very touch centric, which is good for car computing, existing apps will not scale well to the standard 7” car computing touchscreen. Which means that any software which would be properly useful in a car, on the Android OS, would have to be built exclusively for that purpose. That seems fine on the surface, but it excludes you from the primary Android market, and creates a new niche which may be extraordinarily small, if it takes off at all.

Level of interactivity

Android expects the user to be engaged and actively interacting with the device. This is a reasonable expectation for a mobile phone that you’ll use to play games, use social networking, web browse, etc. But the car computing use case is VERY different, you want to be able to interact with the device with as little cognitive overhead as possible. This is the reason that the use of mobile phones is being outlawed in many states. So again, you’d have to build something exclusively for car computing, on a platform with an entirely different focus.

Pay it back

Android is based on Linux, but it is not true to the spirit and intent of Free Open Source Software (FOSS). Android has reinvented the wheel many times over, and created subsystems which do (almost) the same thing as a similar subsystem already available in Linux, but in a way that’s only useful on a mobile device. Fundamentally, this makes sense, since the problems they’re trying to solve are rather specific to the application. The problem though, is that none of that work done to “enhance” the FOSS software upon which Android is based is of any use to anyone but the Android team.

This problem would be even further compounded if Android OS were used in the context it was brought up by @sandbarmark for #carchat. [twitter_status_by_id id=“13882779065”]

Can you imagine? Let’s suppose GM does decide to adopt the Android OS, and turn it into their own version of Ford’s wildly popular Sync entertainment system. Chances are they’re going to take the already very device specific Android OS, take a few key pieces from it, and write their own application suite on top of it. This would mean MORE duplicated work which would likely prove to be only useful to GM, and not to the Android community at large, or the Linux community to which it owes it’s existence!

My Conclusion

While at first blush, the idea of an “open” mobile OS powering in car entertainment devices seemed very attractive I don’t think it’s the answer. If GM is going to pursue building some software and hardware for in car entertainment, they should take a serious look at simply building on top of Linux, rather than having to work around the mobile-device-centric limitations of Android. Plus, if they did it right, the entire Linux community could benefit from their efforts.