[Commotion-dev] convincing google to support ad-hoc mode

[Hans-Christoph Steiner]

On May 25, 2012, at 5:30 PM, L. Aaron Kaplan wrote:

> 
> On May 25, 2012, at 8:07 PM, Hans-Christoph Steiner wrote:
> 
>> 
>> On May 24, 2012, at 10:25 AM, L. Aaron Kaplan wrote:
>> 
>>> 
>>> 
>>> On May 24, 2012, at 2:39 PM, Sascha Meinrath wrote:
>>> 
>>>> Perhaps we should see about setting up a meeting with the Android development
>>>> team management to ensure they understand the importance of ad hoc mode for the
>>>> open source movement?
>>>> 
>>> YES!!
>>> 
>>> That would be a great contribution. 
>>> 
>>> Aaron.
>>> 
>>> 
>>> PS: while you are at it: please tell them that for good meshing we need the layer 2 parameters (RSSI, noise, SNR, biterror rate, etc.) as data readable from /proc or debugfs
>> 
>> Are these not normally exposed in Android?
> 
> I don't know about Android specifically but in Linux there is a great need to have a uniform/standardized way to read out this information. Often enough, when/if you can get these values they are not comparable. Or for example one card will show signal strength but no noise value.
> 
> FYI:
> Currently Henning is also implementing the DLEP draft (https://datatracker.ietf.org/doc/draft-ietf-manet-dlep/) which aims at bringing radio layer information to the mesh router in a standardized way.

Ah ok.  My guess is that Google is very unlikely to tackle this problem.  From what I've seen, they want to stay out of dictating how Android vendors should build their kernels, let alone work in kernel interfaces like this.  My guess is that this effort should be aimed at the Linux kernel people.

.hc