Without complicating things by requiring a dew point to be keyed in (which you may not know at any given altitude at any given moment), it assumes a certain humidity, though I'm not sure of the value. A cursory look shows that it may not be the "worst case" possible spread, which is when is at 100% rel humidity (no temp/dewpoint spread). We're going to look into whether it makes sense to adjust the algorithm for DA.