# DateDifference isn't consistent when giving result in years

This isn't really causing me a problem at the moment, but it's weird. Is this documented anywhere?

When I ask for the difference in days and then in years between April 1, 2015 and October 1, 2015 the years comes out to be exactly .5, whereas I think it should be .50137.

If I use January 1 & July 3, which is the same number of days, I do get .50137.

DateDifference[{2015, 4, 1}, {2015, 10, 1}, "Day"]
%/Quantity[365. , "Days"/"Years"]
DateDifference[{2015, 4, 1}, {2015, 10, 1}, "Year"]

183 days
0.50137 years
0.5 years

DateDifference[{2015, 1, 1}, {2015, 7, 3}, "Day"]
%/Quantity[365. , "Days"/"Years"]
DateDifference[{2015, 1, 1}, {2015, 7, 3}, "Year"]

183 days
0.50137 years
0.50137 years

The reason for the result of

DateDifference[{2015, 4, 1}, {2015, 10, 1}, "Year"]

(* Quantity[0.5, "Years"] *)

is that 2016 is a leap year, so there are 366 days in the year between {2015, 4, 1} and {2016, 4, 1}. Therefore we get 183/366 or exactly 1/2 back.

Note also the documented DayCountConvention option,

DateDifference[{2015, 4, 1}, {2015, 10, 1}, "Year", DayCountConvention -> "Actual365"]

(* Quantity[0.50137, "Years"] *)

and the Wikipedia article on the same topic.