When it is flat lying on the table, it is supposed to show something close to X:0, Y:0, Z:1000. That is: only Z acceleration (earth gravity).
I dont think it has a missing library, as I did not use any external libraries beside the normal python internals.
You can start it on the Xterm with "python /opt/ShipRolling/main.py".
Please let me know if you see an error there.
You should see tons of lines like "Time since last refresh: 248 ms".
Or when you run it with "python /opt/ShipRolling/main.py debug", it will also show the acceleration values.

Can you also use the built in "Write to file" function and send me the file?
Out of that it should be easy to see if doesnt work properly.

I created a log with the debug parameter and exported some values to the csv. (needed a minute to find the file)
The log and the csv file are not from the same session. In the beginning I moved the Device around about 90degree in each direction and after the half I put the N900 flat on the table. I've looked also for a minute in your sourcecode. If the smoothing algorythm is working (have not checked in detail) on your device I don't know why this shouldn't work on my device. I use exactly the same code at my own python script to read from the sensor.
Hmm... strange.

Hi Helex
I had a look on your files.
The csv file looks quite ok, how ever it shows that the phone was lying on the table all the time (angles almost 0) and X and Y almost 0 and Z almost 1000.

The Xterminal log I never really used for that kind of debuging (more for timing issues), so dont spend to much time with it.
All I do is to read the sensor 50 times and then take the average of it as smooting.

What is shown on your graph when you move and tilte your phone?
Can you add a screen shot?
I also would suggest to check with some other acceleration tools like libqacceltests to make sure every thing is correct in your phone.

Here you have a Screenshot. I started your application with opened Flipstand, after 60 seconds I began to move the Device around for about 20 seconds (hold it vertical for 5 seconds) and the rest of the time it layed flat on the table.

I know, you don't belive me. And your source seems also very fine to me.

It works proper on other applications. At example AccDisplay. So I guess the problem is the smoothing algorithm. It seems to me there is no smoothing working at my device. The Diagramm looks very noisy to me. (if I had it flat on the ground, why are there such periodic sparks?)

Hi helex

I realy am clueless why it does not work with your phone.
I made also anothe rapplication with a similair algorithm: SleepAnalyser.
It can record your movement during your sleep.
Could you run this one sometimes, while the phone lies on a table?
I would like to know if it happens there as well.
Several othe rpeople use those apps, but none seems to have this problem.

ShipRoing has reached the level to join Extras-Testing.
Please test version 1.0.7 and vote for it under http://maemo.org/packages/package_in...rolling/1.0-7/

Thank you

What's the difference to sleepy?
I use it for several weeks now, the alarm time is pretty good.


Dan

Hi Threema

I guess you wanted to comment sleepanalyser instead shiprolling.
sleepy just uses a very basic algorithm to detect your movement. I tried to improve that and also to visualise it.Please be aware that sleepanalyser does not yet have an alarm function.

@caco3

What do you think about to build an aplication that trigger an alarm when 'someone' take your phone based on acelerometers?

@BLC: I think I already saw something like that in the ovi store.
Else it also can be done with my SleepAnalyser.
Just have a look on it. Set the alarm window to maximum and the alarm time into the near future.