I'm not entirely sure what aspects of car computing they're talking about here, but from what little meat there is it sounds like they're talking about entertainment and some body control (heat/ac/radio/nav/dvd), not the (to me) more interesting engine/transmission/traction control units.
If my assessment is correct, they're missing a simpler and better opportunity here: the ability to standardize physical/electrical interfaces for these subsystems.
It will be great that they can save 15% (maybe) of their development costs on their nav system interface, but they're still going to be behind the curve when it comes to innovation in this growing space. For example, if the latest craze in in car entertainment were to be a docking station for your Game Boy, they're no further ahead than they were. If the standards were open, the aftermarket could sell a device that would plug in to the harness and integrate seamlessly.
The iPod was a lesson here, and I'm not convinced that anyone sees it as an instance of a generic class of problem -- only a technology that they had to eventually figure out how to cooperate with.
So that got me thinking... an interesting question to ask would be: how do you implement "ship early, ship often" for the software running a car. That mantra has been one of the key facets of open source culture. We ship beta versions, for the hackers who care, and the hardcore folks can just svn checkout the trunk. Bugs get caught early, blah blah blah.
Obviously, it'd be a hard pill to swallow-- just dropping beta software into your shiny new BMW.
So if two BMWs from different product generations could share mostly the same API, standardized protocols, etc., well then we could test stuff with the old E46 before dropping it into the new E92...
Legacy form factor mules with advanced technology in them.
Phase two: Performance tuning shops often pick up early versions of cars to start testing, looking for saleable mods, and adapting their lines to the new products. I would expect we'd start seeing reflashes for systems within a few months of a vehicle release
Phase three: As platform N depreciates and platform N+1 goes into production, N-class cars start falling into the hands of people more willing to cut them open and hack (track/hobby cars). If N and N+1 run the same code (project/source level, not binaries), features could be committed by individual hackers and could be picked up by manufacturers as desired.
Of course this is all predicated on this release being a real, public, open source project. If their desire/intent is to create an open source platform so that 2-5 manufacturers can share IP, it's possible that we'll never see the infrastructure to flash things ourselves, in which case this is all a big fat stinky dead end. :/
Well its pretty simple really. How hard would it be for them to redesign their navigation systems to include a usb port? Then the user would go to apps.bmw.com download your app and bam.
The reason I'm thinking of them using the navigation system is because you'll need a bigger screen to make the whole app thing work.
As far as compatibility between generations I wouldn't worry too much about that. Most likely on top of an API they'd release an emulator.
I don't think they stress over the the APIs and standardized protocol stuff, its much more consuming tuning the system to provide optimal performance than it is to swap out a computer. A modern car now has tons of sensors and it takes a lot of testing and fine tuning to get every last bit of performance out at the 'ring.
We're going to be authoring our own software. We'll open source it once we're done, but we're not going to let partnerships slow us down in this phase.
If my assessment is correct, they're missing a simpler and better opportunity here: the ability to standardize physical/electrical interfaces for these subsystems.
It will be great that they can save 15% (maybe) of their development costs on their nav system interface, but they're still going to be behind the curve when it comes to innovation in this growing space. For example, if the latest craze in in car entertainment were to be a docking station for your Game Boy, they're no further ahead than they were. If the standards were open, the aftermarket could sell a device that would plug in to the harness and integrate seamlessly.
The iPod was a lesson here, and I'm not convinced that anyone sees it as an instance of a generic class of problem -- only a technology that they had to eventually figure out how to cooperate with.