Once you put them on pipes, you have to recalibrate them anyway, so what is the point to have them calibrated already.
We have found a few problems with this approach. First, there is a small bias introduced from mounting them on the pipe surface. However, because the bias applies to both the EWT and LWT, its affect will be small (less than 5% error in delta T), if properly installed -- using the methods similar to those described by Bill. The sensor offset is a different source of error and, because it is temperature dependent (as shown in DS18B20 datasheet and App note cited above), both the slope and offset need to be incorporated into calibration. The final challenge with calibration in the field is that temperature probes that are available for P/T ports are typically either very expensive or not very accurate (in spite of some manufacturers claims).
Here is a link to a fact sheet we've put together that summarizes these issues. http://groundenergysupport.com/wp/docs/datasheets/GxTracker_DataSheet_OnPipeTemp_rev1.1.pdf