"Guess why they name division "floating point operation"?"
Got that and sorry I was not clear, but intend to ask "what is the precision and scale of the results?" Finally found the answer in BOL but I am rephrasing slightly.
Given expression e1, with precision p1 and scale s1, and expression e2, with precision p2 and scale s2, the result of e1 divided by e2 is:
Precision is p1 - s1 + s2 + max(6, s1 + p2 + 1)
Scale is max(6, s1 + p2 + 1)
SQL = Scarcely Qualifies as a Language