If you haven't read my first post on Project Thermostat, you might want to check it out first.
I decided to just go out and buy the exact same thermostat we had before on the Dutch ebay variant. A second hand one that looked really battered only cost me 15 euros. I took out the damaged print (leaking battery stripped away some of the conducting lines) and replaced it with the battered one.
Problem solved! But I did have a nice time looking for a self-built alternative. It should be really simple, but would costs at least as much as a ready built one. I'm wondering how much of the costs of these little devices goes into certification and standardization trials, patent costs etc. It must be a lot, because I can, with simple devices such as the .NET Gadgeteer create one with a material cost that is around as much as what a ready made device costs. If you add cool features such as wireless networking, usb support, remote access and others, then I can put together the hardware way below the costs of the alternatives offered in the market right now.
If anyone takes up an open Thermostat project, I'd like to dump some of my ideas ;). Until then I'll keep buying secondhand Chronotherm 4's.