Why not measuring the outside ambient air? The surrounding air is the most important game of all to measure -- ambient air ultimately is controlling everything else in the flight envelope.
The point is that to get good results from solid state gyros and accelerometers, you need to compensate for temperature. Hence one temp probe per sensor as mentioned in this post:
http://www.pprune.org/rumours-news/3...ml#post5007882
What matters is the temperature of the sensor itelf, not the outside temperature or even the air temperature inside the ISIS box.
Semiconductor chips are pretty good at measuring themselves. A patch of calibrated diodes somewhere on the chip and a mux into a ADC is all one needs for zero-pincount internal temp precison measurement, suitable for whatever corrections might be desired for precision calibration. Pins are more valuable than gold on most chips.
That really depends on the application (pin count issue). I doubt this is a highly integrated system using custom ICs. The main goal would be accuracy.