Self-heating current limit for temp sensor

http://www.sparkfun.com/commerce/product_info.php?products_id=9438

It's basically a diode temperature sensor. I'm going to be using one of those in an air temperature chart recorder, with no heatsink. The datasheet is frustrating because it doesn't say how much current is acceptable, it only says to minimize it, but warns that you need 'enough to drive the sensor and the calibration pot'...whatever that means.

I need to figure out what size resistor to use. It will be hooked up to 5v and the sensor drops 3v so that means 2v will be dropped over the resistor so if I'm thinking right, a 1k resistor would be 2ma. Do you think 2ma is a good value? How could too little be a problem; why shouldn't I just use like a megaohm?

Too little current is a problem because it's not a diode temp sensor. It is a circuit which just happens to act (more or less) like a diode sensor. The 1mA that all the specs are listed at is probably a good value. You could of course put it in still air of a known temp and try different currents to see how much self-heating error is produced.

or maybe you could ask sparkfun, I hear rumor they support the products they sell (though I have not seen it)