There was quite a lot of interest in my gas measuring project using a mobile phone. Thanks for that. For all who requested the source: I packed it together and wrote a small readme. You can download it now.
There were doubts about the accuracy and if it will work at all. It is running for over a week now without any crash or other problems. It measured a gas usage of 14.81 m³ while the gas meter is showing 13.85 m³. That is a difference of 6.9%. Not perfect, but i can live with that and can try to reset the “burning duration -> volume” factor to reflect the difference.
I also have first usage graphs:
Gas usage for September 2008
Day of month on x axis and used energy in kWh on y axis.
Does it mean a lot: No, not yet. Not enough days and also to much playing around with settings (temperature on thermostat in the kitchen, setting on the radiators).
One interesting fact is that it will always do exact 11 “burning sessions” every hour as long the temperature in the room with the thermostat has not reached the target temperature. But the duration of each of those sessions depends on the real needed energy from the radiators. Even if they do not request any energy it will do 11 burning starts but all of them very short.
That means: At least here is quite some room for optimization. I need to make sure that it only starts when at least one room requests energy. That is quite hard to achieve because there is only a thermostat for controlling it in one room. 
Here a chart for an average day:
Gas usage graph for 2008/09/21
Hour of the day on x, burning duration in seconds on y
The two peaks are from the warm water heating. That starts about two times a day and takes about 250 seconds or 0.7 kWh each, means about 2 kWh (ca. 0.15 EUR ) a day for warm water.
Anyone still reading? 
