I don't think you could use rctime or pot to accurately measure temperature. You'll find that, in order to get some decent resolution; i.e. 0.5° steps - you'll need an ADC with 10 bit capability. Many senors, LM335 for example, induce 10mV changes p/ degree. So, if we have a 10 bit ADC we can detect changes as low as 5mV. (5 /1024) = 4.88mV actually.