GSoC 2008

[Crane, Matthew]

> But as we only can choice one of them for this application, you should
> be prepared for other applications, too.

Yea, whatever API into the accelerometer is made should return some form
of condensed data but not necessarily be tied to the idea of gestures.  

Different applications may want the data at the same time.  

E.g. background car crash detector (that dials out and knows when it's
on the road) working concurrently with gestures for answering phone etc.

Maybe the sort of thing where an api would allow an app to register a
set of gesture, defined mathematically, and only one system process
polls for matching events.  Doesn't sound like multiple processes
polling the data, or even processing the gestures, would work as nicely.

Matt


-----Original Message-----
From: community-bounces at lists.openmoko.org
[mailto:community-bounces at lists.openmoko.org] On Behalf Of Stefan
Schmidt
Sent: Tuesday, March 25, 2008 10:36 AM
To: community at lists.openmoko.org
Subject: Re: GSoC 2008

Hello.

On Mon, 2008-03-24 at 19:02, Niluge KiWi wrote:
> 
> I'm interested in the accelerometers features [1]: Recognising
gestures
> is a really important part of the interface between the user and the
> phone.

Seems this ideas gets the interest of a lot people. Nice. :)


> With the two accelerometers in the FreeRunner, I think we can
recognise
> lots of gestures, not only simple ones like a "click" (which is
already
> recognised by the accelerometers used in the FreeRunner). The main
> difficulty is probably to extract the useful data from the gestures
> noise : calibration may take time. The goal is to have an almost
> pre-calibrated library (an idea from the wish-list in the Wiki is to
> allow the user to record its own gestures, but I think it's not easy
to
> do it simple for the end-user).

Letting the user add new gestures is a key feature IMHO. Also letting
them combine different gestures to new ones. We should make it easy
for people beaing creative with this. That's where innovation can
start. :)

If we can have a preset of already known gestures shipped with the
device, great.

> I'm also interested in working in the ambient noise detection in
second
> choice.

Also interesting. What I never understand completely is what kind of
cool stuff we can do with this. I mean detecting the ambient volume
level and adjust the ringing, etc is nice, but can we do more with it?
Fancy things like detect if we are in a car or plane and react
accordingly?

regards
Stefan Schmidt