Thanks Daniel. It's taking shape. Interesting to see how it performs.
The latest is a screen shot of the temperature sensor readout, polled every second. The sensor is reasonably stable with the odd spike +/-, > 0.3 degrees from average. Polling at longer intervals might be better for the control system, given that temperature doesn't change rapidly.
Here
EDIT: I should mention that checked against two separate electronic thermometers at 27C and 28.5C, accuracy is not too bad.
More important is temperature control, so I need to decide which parameter to read for the best sensitivity.