adf.ly

Tampilkan postingan dengan label BMW. Tampilkan semua postingan
Tampilkan postingan dengan label BMW. Tampilkan semua postingan

Rabu, 30 September 2009

GenIVI – “Generally Expected Not for In Vehicle Infotainment”

Last week at the Intel Developer’s Forum (IDF) in San Francisco, Intel CEO, Paul Otellini announced in his keynote that Intel’s Atom processor would be used in upcoming 2012 Model Year BMW and Daimler vehicles. The OS to be used? QNX Neutrino RTOS. What I find interesting is:


So at the end of the day team GenIVI didn’t give itself the win in stacked competition? You have to wonder why that might have been? Perhaps we should look to Linus Torvalds’ recent commentary at LinuxCon where he stated that the Linux kernel has become "huge and scary" and that it isn't "the streamlined, hyper-efficient kernel I envisioned when I started writing Linux."

Perhaps GenIVI should consider its OS strategy. . . QNX might be a good choice

Rabu, 11 Maret 2009

CeBit 2009 - Infotainment Comparison

Wow! CeBit exhibition was down almost 30% from last year and nearly 50% from its peak. And, though I heard official numbers put attendance down only 20% from last year, it seemed to me more like 50%. With the number of backpacks that seemed to emerge on Thursday and Friday, I wonder if attendance was inflated with students.

All of that being said, automotive was well represented with booths and / or attendance by GM, Audi, VW, BMW, PSA and others. Tier 1s, Continental and Harman Becker, each had demonstration systems . . . with the HBAS system far more responsive and impressive (see video comparison)

I would have compared the GenIVI Linux based solution, but as you can see from the HBAS video, Intel is still pushing the QNX based demonstrations as the Linux system doesn't yet exist. ;)

Selasa, 10 Maret 2009

When a license is the "single point of failure"

Last week, at CeBit, I had the opportunity to witness the launch of the GenIVI consortium. Graham Smethurst, of BMW, did an excellent job summarizing the challenges facing automotive OEMs and Tier 1s that led to the creation of GenIVI. The theory is that you can plug-in any conforming middleware (speech rec, navigation engine, etc.) into the GenIVI framework. The ultimate goal of GenIVI being to create standard interfaces to reusable modules. I walked away struggling with only one point: his contention that everything in GenIVI was created to remove any "single point of failure".

The reason I struggled with this point was because of the use of a single OS, Linux, for GenIVI. While there might be many distributions of Linux they all share one thing in common the GNU Public License (GPL). If Linux is the only OS for GenIVI, then it stands to reason that GPL then becomes THE single point of failure. (see my thoughts on GPL in automotive)

If it were me, I would have built GenIVI on an industry standard API like POSIX rather than a specific OS. In this way even the OS licensing could be varietal (GPL, LGPL, BSD, etc.)