Car software has three huge challenges most other devices can be forgiven on more readily: reliability, speed and integration.
Reliability
Car software needs to be bug-free, or darn close to it. A person can’t be driving down the road when suddenly their radio volume knob stops working, or their navigation system crashes. For this reason, auto manufacturers have a strong impulse to build in-house, where they have complete control of the end-to-end experience. Outsourcing, or investing in a startup is too high a risk (on paper) when it comes to driver safety. Could you imagine driving a car with version 1.0 of an Android-based system developed by a mature company, let alone a startup? I’d be terrified.Speed and Integration
When you turn on your car, the software needs to be available immediately, or within a tenth of a second. No functionality can be locked behind any kind of “Loading…” screen. The radio needs to switch on right away and navigation needs to be available, at the touch of a button. For true performance, the UI and software for cars needs to be one step above hard-coded.If a software company is focusing on modular software for cars, they would need to build software specifically designed for each manufacturer to be fast and integrated enough. In that scenario, there’s no business model that makes sense. Imagine setting out to reinvent Mercedes’ car computers. Now, imagine them agreeing to a deal with your company without seeing a flawless prototype. It’d never happen.
So, to answer the question posed above, existing as an external company providing software for cars is a pretty tough business model to pitch. The only way to yield a truly great UI is to hire amazing designers in-house. Since the software doesn’t necessarily sell the car (for most people), the company’s investment follows in kind. I, too, hope this changes in the future.
Random idea: Maybe an HTML5 UI would allow for re-use, despite a completely different low-level code base.
I’m not arguing in the slightest that this is something car companies should outsource, or that it’s an opportunity for some startup to disrupt, I’m arguing that they have no excuse not to get their shit together, and that it’s especially vexing to see something otherwise exquisitely designed have such shitty UI (both hardware AND software.)
There’s nothing about car software’s need for “reliability, speed and integration” that’s preventing it from being good. In fact, those are all reasons it should be orders of magnitude better than the typical embedded device. The human part of it can absolutely be separated from the mechanical part of it, the processor handling my brakes should not be handling my radio and GPS. You know what you see on the screen every single time you start a BMW? A fucking end-user license agreement that you have to click to accept. HTML5 is not going to correct poor taste.
We’re five years post-iPhone, people have had a taste of the good stuff and there’s no going back. Cars themselves are becoming increasingly commoditized and there’s a massive opportunity to differentiate based on user interface. Instead we get shit like this.
(Source: jstn)
loading…
